The ongoing COVID-19 situation, declared as pandemic on 11 March 2020 by WHO, has adversely affec... more The ongoing COVID-19 situation, declared as pandemic on 11 March 2020 by WHO, has adversely affected people's lives around the globe since its beginning at Wuhan, China, in December 2019. According to the Centers for Disease Control and Prevention (CDC), the national public health organization of the US, both type 1 or type 2 diabetes mellitus are risk factors for COVID-19. In Southeast Asia (SEA), around 90 million people have been reported to suffer from diabetes mellitus, which is characterized by deregulated use of blood glucose by the body 1 . The Mayo Clinic, a non-profit medical center in the US, suggested that regular exercise, weight loss, healthy eating, and anti-diabetic medication or insulin, help to manage the disease. The lockdown situation due to COVID-19 has compelled people around the globe to stay at home, preventing them from going to parks or a gym. Most people in the lowermiddle-income countries (LMIC) in SEA, like Bangladesh or India, which are densely populated, cannot afford spacious living places due to their poor financial state and restricted movement in the city areas 2 . In general, most diabetes patients in these countries could not go outside their homes for exercise or walking during lockdown due to movement restrictions, and most of them have had to stay in confined accommodation. Moreover, a gym at home is unavailable for most people living in LMICs. These barriers to physical exercise may have prompted uncontrolled glycemic situation throughout the pandemic period. A study in India reported HbA1c increased by 2.26% from baseline after 30 days of lockdown and 3.68% after 45 days of lockdown 3 . According to research, an increase of 0.51% in the HbA1c level was found in an Indian studied
Background: E-learning is making education globally and conveniently attainable with the delivera... more Background: E-learning is making education globally and conveniently attainable with the deliverance of advanced technology. However, this mode of academia is still not commonly practiced locally. Thus, the study aimed to investigate technological availability, usability, and association to university students' perceived stress due to e-learning curriculum. Methods: A cross-sectional study commenced among Bangladeshi university students enrolled in the e-learning curriculum. A total of 1162 university students were included. The main explanatory variables were related to the availability of technology and the use of technology. The outcome variable was perceived e-learning stress. In statistical analysis, p-value < 0.05 was considered statistically significant with a 95% confidence interval. Results: In this study, lack of technological availability and usability were associated with higher level of perceived e-learning stress. Being female, living in rural areas, and outside...
International Journal of Sustainable Agricultural Research, 2021
Association of sustainability agriculture and farming practices is somehow closely connected. The... more Association of sustainability agriculture and farming practices is somehow closely connected. There are necessary different farming practices for both adjusted and unadjusted PFSI measurement. The study observes practices of paddy farming and if farmers are practicing agriculture sustainably by estimating PFSI in three villages of Gutudia union. The objective is to spot the present agricultural practices and accessible sustainable practices, to examine the sustainability degree at field beneath the present paddy farming systems using PFSI and additionally to identify recommendations. The unit of analysis is 50 farmers and measured on a scale of 0 to 100 and also through Saltiel, Bauder, and Palakovich (1994) index. The results discovers that the average sustainability level which is presumably quite unsustainable this shows the necessity for more extension of correct practices. Chi-square analysis shows that the level of farmers’ awareness toward sustainable agriculture and positive...
Goal: Although photoplethysmogram (PPG) and electrocardiogram (ECG) signals can be used to estima... more Goal: Although photoplethysmogram (PPG) and electrocardiogram (ECG) signals can be used to estimate blood pressure (BP) by extracting various features, the changes in morphological contours of both PPG and ECG signals due to various diseases of circulatory system and interaction of other physiological systems make the extraction of such features very difficult. Methods: In this work, we propose a waveform-based hierarchical Artificial Neural Network - Long Short Term Memory (ANN-LSTM) model for BP estimation. The model consists of two hierarchy levels, where the lower hierarchy level uses ANNs to extract necessary morphological features from ECG and PPG waveforms and the upper hierarchy level uses LSTM layers to account for the time domain variation of the features extracted by lower hierarchy level. Results: The proposed model is evaluated on 39 subjects using the Association for the Advancement of Medical Instrumentations (AAMI) standard and the British Hypertension Society (BHS) ...
Lung cancer is one of the most severe and widespread that constitutes a major public health probl... more Lung cancer is one of the most severe and widespread that constitutes a major public health problem and has a high mortality rate. In this regard, proper segmentation of lung tumor from X-ray, Computed Tomography (CT scan) or, Magnetic Resonance Imaging (MRI) is the stepping stone towards achieving completely automated diagnosis system for lung cancer detection. With the advancement of technology and availability of data, the valuable time of a radiologist can be saved using computer tools for tumor segmentation. In this work, we present a data driven approach for lung tumor segmentation from CT scans by using Recurrent 3D-DenseUNet, a novel fusion of Convolutional and Recurrent neural network. Our approach is to train this network using image-volumes with tumor only slices of size (256 X 256 X 8). A data-driven adaptive weighting method is also used in our approach to differentiate between tumorous and non-tumorous image-slices, which shows more promise than crude intensity thresho...
The present study attempts to determine the productivity, profitability and resource use efficien... more The present study attempts to determine the productivity, profitability and resource use efficiency of four promising spices crops such as garlic, chilli, ginger and turmeric. The data were collected from 480 farm households in the crop year 2010-2011. Productions of all the crops were profitable as estimated by net returns and benefit cost ratios. Functional analyses showed that farm size, seed, inorganic and organic fertilisers, cost of power tiller and draft power, irrigation, education, farming experience and training had positive impact on the production of spices. Increasing returns to scale prevailed in the production process for garlic, chilli, ginger whereas constant returns to scale prevailed for the production of turmeric. All the models used fitted well to analyse the selected data for all crops. Small farmers were more efficient for garlic production only whereas the large farmers were more efficient for other spices crops. More educated and more experienced farmers wer...
The present effort was to obtain extracts from various fruit by-products using three extraction s... more The present effort was to obtain extracts from various fruit by-products using three extraction systems and to evaluate their polyphenolic content, antioxidant, and α-glucosidase inhibition activity. The fruit by-products were pre-processed by washing, drying, and milling methods to produce the powder. The powder samples were used to obtain extracts using pressurized hot-water (PHWE), enzyme-assisted (EnE) and organic solvent extraction (OSE) systems. The total phenolic content (TPC), total flavonoid content (TFC), antioxidant and α-glucosidase inhibition activity in all samples were assessed by Folin-Ciocalteu, AlCl3 colorimetric, DPPH· & ABST·+ and α-glucosidase inhibitory methods. The results showed that the extracts of peel, seed and other by-products exhibited outstanding TPC, TFC, and strongest antioxidant and α-glucosidase inhibition activity, eventually higher than edible parts of the fruits. For instance, the highest TPC among the peels of various fruits were in mango peel (in all cultivar) followed by litchi peel, banana peel cv. sagor, jackfruit peel, pineapple peel, papaya peel, banana peel cv. malbhog and desi on average in all tested extraction systems. PHWE system yielded significantly (p < 0.05) higher TPC and TFC than other extraction systems. In case of misribhog mango variety, the TPC (mg GAE/g DM) in peels were 180.12 ± 7.33, 73.52 ± 2.91 and 36.10 ± 3.48, and in seeds were 222.62 ± 12.11, 76.18 ± 2.63 and 42.83 ± 12.52 for PHWE, EnE and OSE respectively. This work reported the promising potential of underutilized fruit by-products as new sources to manufacture ingredients and nutraceuticals for foods and pharmaceutical products.
Background Mucormycosis, a severe fungal infection, is an emerging public health concern during t... more Background Mucormycosis, a severe fungal infection, is an emerging public health concern during the COVID-19 pandemic. This study aimed to investigate the perception of mucormycosis among Bangladeshi healthcare workers. Results An exploratory cross-sectional study was carried out among the Bangladeshi healthcare workers from May 25, 2021, to June 5, 2021. The study found 422 responses from the healthcare workers of Bangladesh. Among the respondents, nearly half of them (45.26%) were doctors (n = 191). This study explored that the healthcare workers’ mucormycosis perception scores were significantly associated with their age, gender, profession, monthly income, marital status, job type, and death of friends and family members due to COVID-19. Conclusions This study emphasized the healthcare workers’ mucormycosis perception along with other associated factors. The findings could help policymakers to mitigate mucormycosis and related infectious diseases emergencies in the post-COVID-19...
Proceedings of the 3rd IEEE International Symposium on Signal Processing and Information Technology (IEEE Cat. No.03EX795)
Department of Electrical and Electronic Engineering, Bangladesh University of Engineering and Tec... more Department of Electrical and Electronic Engineering, Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh ... ABSTRACT This paper illustrates the superiority of using time-frequency varying averaging factor to estimate the a priori SNR far
Source camera model identification (CMI) and image manipulation detection are of paramount import... more Source camera model identification (CMI) and image manipulation detection are of paramount importance in image forensics. In this paper, we propose an L2-constrained Remnant Convolutional Neural Network (L2-constrained RemNet) for performing these two crucial tasks. The proposed network architecture consists of a dynamic preprocessor block and a classification block. An L2 loss is applied to the output of the preprocessor block, and categorical crossentropy loss is calculated based on the output of the classification block. The whole network is trained in an end-to-end manner by minimizing the total loss, which is a combination of the L2 loss and the categorical crossentropy loss. Aided by the L2 loss, the data-adaptive preprocessor learns to suppress the unnecessary image contents and assists the classification block in extracting robust image forensics features. We train and test the network on the Dresden database and achieve an overall accuracy of 98.15%, where all the test images are from devices and scenes not used during training to replicate practical applications. The network also outperforms other state-of-the-art CNNs even when the images are manipulated. Furthermore, we attain an overall accuracy of 99.68% in image manipulation detection, which implies that it can be used as a general-purpose network for image forensic tasks.
Ultrasound Shear Wave Elastography (USWE) with conventional B-mode imaging demonstrates better pe... more Ultrasound Shear Wave Elastography (USWE) with conventional B-mode imaging demonstrates better performance in lesion segmentation and classification problems. In this article, we propose SHEAR-net, an end-to-end deep neural network, to reconstruct USWE images from tracked tissue displacement data at different time instants induced by a single acoustic radiation force (ARF) with 100% or 50% of the energy in conventional use. The SHEAR-net consists of a localizer called the S-net to first localize the lesion location and then uses recurrent layers to extract temporal correlations from wave patterns using different time frames, and finally, with an estimator, it reconstructs the shear modulus image from the concatenated outputs of S-net and recurrent layers. The network is trained with 800 simulation and a limited number of CIRS tissue mimicking phantom data and is optimized using a multi-task learning loss function where the tasks are: inclusion localization and modulus estimation. Th...
Camera model identification has earned paramount importance in the field of image forensics with ... more Camera model identification has earned paramount importance in the field of image forensics with an upsurge of digitally altered images which are constantly being shared through websites, media, and social applications. But, the task of identification becomes quite challenging if metadata are absent from the image and/or if the image has been post-processed. In this paper, we present a DenseNet pipeline to solve the problem of identifying the source camera-model of an image. Our approach is to extract patches of 256*256 from a labeled image dataset and apply augmentations, i.e., Empirical Mode Decomposition (EMD). We use this extended dataset to train a Neural Network with the DenseNet-201 architecture. We concatenate the output features for 3 different sizes (64*64, 128*128, 256*256) and pass them to a secondary network to make the final prediction. This strategy proves to be very robust for identifying the source camera model, even when the original image is post-processed. Our mo...
Recognizing humor from a video utterance requires understanding the verbal and non-verbal compone... more Recognizing humor from a video utterance requires understanding the verbal and non-verbal components as well as incorporating the appropriate context and external knowledge. In this paper, we propose Humor Knowledge enriched Transformer (HKT) that can capture the gist of a multimodal humorous expression by integrating the preceding context and external knowledge. We incorporate humor centric external knowledge into the model by capturing the ambiguity and sentiment present in the language. We encode all the language, acoustic, vision, and humor centric features separately using Transformer based encoders, followed by a cross attention layer to exchange information among them. Our model achieves 77.36% and 79.41% accuracy in humorous punchline detection on UR-FUNNY and MUStaRD datasets – achieving a new state-of-the-art on both datasets with the margin of 4.93% and 2.94% respectively. Furthermore, we demonstrate that our model can capture interpretable, humorinducing patterns from al...
Removing speckle noise from medical ultrasound images while preserving image features without int... more Removing speckle noise from medical ultrasound images while preserving image features without introducing artifact and distortion is a major challenge in ultrasound image restoration. In this paper, we propose a multiframe-based adaptive despeckling (MADS) algorithm to reconstruct a high-resolution B-mode image from raw radio-frequency (RF) data that is based on a multiple input single output (MISO) model. As a prior step to despeckling, the speckle pattern in each frame is estimated using a novel multiframe-based adaptive approach for ultrasonic speckle noise estimation (MSNE) based on a single input multiple output (SIMO) modeling of consecutive deconvolved ultrasound image frames. The elegance of the proposed despeckling algorithm is that it addresses the despeckling problem by completely following the signal generation model unlike conventional ad-hoc smoothening or filtering based approaches, and therefore, it is likely to maximally preserve the image features. As deconvolution...
ESD is an important micro-parameter for ultrasonic tissue characterization. However, classifying ... more ESD is an important micro-parameter for ultrasonic tissue characterization. However, classifying between benign and malignant breast lesions from a broad dataset remains a challenge. In this work, we propose a new technique for the estimation of ESD of breast tissues from the diffuse component of backscattered RF data. This allows us to combine ESD with MSS and other ultrasonic macro-parameters for binary classification of lesions. In order to separate the diffuse component from the coherent component of the backscattered data, EEMD is performed. K-S test is used to automatically select the IMFs responsible for diffuse scattering. To ensure proper minimization of the system effects, a multi-step process is adopted where the RF data is deconvolved and filtered as the first two steps prior to normalization. As the ESD is supposed to have a fairly consistent value over a small tissue subregion, it is estimated, from the slope of the average regression line, computed from the exponentia...
An analysis of the noise effect on the convergence characteristic of the least-mean-squares (LMS)... more An analysis of the noise effect on the convergence characteristic of the least-mean-squares (LMS) type adaptive algorithms for blind channel identification is presented. It is shown that the adaptive blind algorithms misconverge in the presence of noise. A novel technique for ameliorating such misconvergence characteristic, using a frequency domain energy constraint in the adaptation rule, is proposed. Experimental results demonstrate that the robustness of the blind adaptive algorithms can be significantly improved using such constraints.
2002 11th European Signal Processing Conference, 2002
This paper presents an efficient thresholding technique for wavelet speech enhancement. The signa... more This paper presents an efficient thresholding technique for wavelet speech enhancement. The signal-bias compensated noise level is used as the threshold parameter. The noise as well as signal level is estimated from the detail wavelet packet (WP) coefficients in the first scale. Both hard and soft thresholding are applied successively. The regions for hard thresholding are identified by estimating their signal to noise ratio (SNR) in the wavelet domain. Soft thresholding is applied to the rest of the regions. The performance of the proposed scheme is evaluated on speech recorded in real conditions with artificial noise added to it.
The ongoing COVID-19 situation, declared as pandemic on 11 March 2020 by WHO, has adversely affec... more The ongoing COVID-19 situation, declared as pandemic on 11 March 2020 by WHO, has adversely affected people's lives around the globe since its beginning at Wuhan, China, in December 2019. According to the Centers for Disease Control and Prevention (CDC), the national public health organization of the US, both type 1 or type 2 diabetes mellitus are risk factors for COVID-19. In Southeast Asia (SEA), around 90 million people have been reported to suffer from diabetes mellitus, which is characterized by deregulated use of blood glucose by the body 1 . The Mayo Clinic, a non-profit medical center in the US, suggested that regular exercise, weight loss, healthy eating, and anti-diabetic medication or insulin, help to manage the disease. The lockdown situation due to COVID-19 has compelled people around the globe to stay at home, preventing them from going to parks or a gym. Most people in the lowermiddle-income countries (LMIC) in SEA, like Bangladesh or India, which are densely populated, cannot afford spacious living places due to their poor financial state and restricted movement in the city areas 2 . In general, most diabetes patients in these countries could not go outside their homes for exercise or walking during lockdown due to movement restrictions, and most of them have had to stay in confined accommodation. Moreover, a gym at home is unavailable for most people living in LMICs. These barriers to physical exercise may have prompted uncontrolled glycemic situation throughout the pandemic period. A study in India reported HbA1c increased by 2.26% from baseline after 30 days of lockdown and 3.68% after 45 days of lockdown 3 . According to research, an increase of 0.51% in the HbA1c level was found in an Indian studied
Background: E-learning is making education globally and conveniently attainable with the delivera... more Background: E-learning is making education globally and conveniently attainable with the deliverance of advanced technology. However, this mode of academia is still not commonly practiced locally. Thus, the study aimed to investigate technological availability, usability, and association to university students' perceived stress due to e-learning curriculum. Methods: A cross-sectional study commenced among Bangladeshi university students enrolled in the e-learning curriculum. A total of 1162 university students were included. The main explanatory variables were related to the availability of technology and the use of technology. The outcome variable was perceived e-learning stress. In statistical analysis, p-value < 0.05 was considered statistically significant with a 95% confidence interval. Results: In this study, lack of technological availability and usability were associated with higher level of perceived e-learning stress. Being female, living in rural areas, and outside...
International Journal of Sustainable Agricultural Research, 2021
Association of sustainability agriculture and farming practices is somehow closely connected. The... more Association of sustainability agriculture and farming practices is somehow closely connected. There are necessary different farming practices for both adjusted and unadjusted PFSI measurement. The study observes practices of paddy farming and if farmers are practicing agriculture sustainably by estimating PFSI in three villages of Gutudia union. The objective is to spot the present agricultural practices and accessible sustainable practices, to examine the sustainability degree at field beneath the present paddy farming systems using PFSI and additionally to identify recommendations. The unit of analysis is 50 farmers and measured on a scale of 0 to 100 and also through Saltiel, Bauder, and Palakovich (1994) index. The results discovers that the average sustainability level which is presumably quite unsustainable this shows the necessity for more extension of correct practices. Chi-square analysis shows that the level of farmers’ awareness toward sustainable agriculture and positive...
Goal: Although photoplethysmogram (PPG) and electrocardiogram (ECG) signals can be used to estima... more Goal: Although photoplethysmogram (PPG) and electrocardiogram (ECG) signals can be used to estimate blood pressure (BP) by extracting various features, the changes in morphological contours of both PPG and ECG signals due to various diseases of circulatory system and interaction of other physiological systems make the extraction of such features very difficult. Methods: In this work, we propose a waveform-based hierarchical Artificial Neural Network - Long Short Term Memory (ANN-LSTM) model for BP estimation. The model consists of two hierarchy levels, where the lower hierarchy level uses ANNs to extract necessary morphological features from ECG and PPG waveforms and the upper hierarchy level uses LSTM layers to account for the time domain variation of the features extracted by lower hierarchy level. Results: The proposed model is evaluated on 39 subjects using the Association for the Advancement of Medical Instrumentations (AAMI) standard and the British Hypertension Society (BHS) ...
Lung cancer is one of the most severe and widespread that constitutes a major public health probl... more Lung cancer is one of the most severe and widespread that constitutes a major public health problem and has a high mortality rate. In this regard, proper segmentation of lung tumor from X-ray, Computed Tomography (CT scan) or, Magnetic Resonance Imaging (MRI) is the stepping stone towards achieving completely automated diagnosis system for lung cancer detection. With the advancement of technology and availability of data, the valuable time of a radiologist can be saved using computer tools for tumor segmentation. In this work, we present a data driven approach for lung tumor segmentation from CT scans by using Recurrent 3D-DenseUNet, a novel fusion of Convolutional and Recurrent neural network. Our approach is to train this network using image-volumes with tumor only slices of size (256 X 256 X 8). A data-driven adaptive weighting method is also used in our approach to differentiate between tumorous and non-tumorous image-slices, which shows more promise than crude intensity thresho...
The present study attempts to determine the productivity, profitability and resource use efficien... more The present study attempts to determine the productivity, profitability and resource use efficiency of four promising spices crops such as garlic, chilli, ginger and turmeric. The data were collected from 480 farm households in the crop year 2010-2011. Productions of all the crops were profitable as estimated by net returns and benefit cost ratios. Functional analyses showed that farm size, seed, inorganic and organic fertilisers, cost of power tiller and draft power, irrigation, education, farming experience and training had positive impact on the production of spices. Increasing returns to scale prevailed in the production process for garlic, chilli, ginger whereas constant returns to scale prevailed for the production of turmeric. All the models used fitted well to analyse the selected data for all crops. Small farmers were more efficient for garlic production only whereas the large farmers were more efficient for other spices crops. More educated and more experienced farmers wer...
The present effort was to obtain extracts from various fruit by-products using three extraction s... more The present effort was to obtain extracts from various fruit by-products using three extraction systems and to evaluate their polyphenolic content, antioxidant, and α-glucosidase inhibition activity. The fruit by-products were pre-processed by washing, drying, and milling methods to produce the powder. The powder samples were used to obtain extracts using pressurized hot-water (PHWE), enzyme-assisted (EnE) and organic solvent extraction (OSE) systems. The total phenolic content (TPC), total flavonoid content (TFC), antioxidant and α-glucosidase inhibition activity in all samples were assessed by Folin-Ciocalteu, AlCl3 colorimetric, DPPH· & ABST·+ and α-glucosidase inhibitory methods. The results showed that the extracts of peel, seed and other by-products exhibited outstanding TPC, TFC, and strongest antioxidant and α-glucosidase inhibition activity, eventually higher than edible parts of the fruits. For instance, the highest TPC among the peels of various fruits were in mango peel (in all cultivar) followed by litchi peel, banana peel cv. sagor, jackfruit peel, pineapple peel, papaya peel, banana peel cv. malbhog and desi on average in all tested extraction systems. PHWE system yielded significantly (p < 0.05) higher TPC and TFC than other extraction systems. In case of misribhog mango variety, the TPC (mg GAE/g DM) in peels were 180.12 ± 7.33, 73.52 ± 2.91 and 36.10 ± 3.48, and in seeds were 222.62 ± 12.11, 76.18 ± 2.63 and 42.83 ± 12.52 for PHWE, EnE and OSE respectively. This work reported the promising potential of underutilized fruit by-products as new sources to manufacture ingredients and nutraceuticals for foods and pharmaceutical products.
Background Mucormycosis, a severe fungal infection, is an emerging public health concern during t... more Background Mucormycosis, a severe fungal infection, is an emerging public health concern during the COVID-19 pandemic. This study aimed to investigate the perception of mucormycosis among Bangladeshi healthcare workers. Results An exploratory cross-sectional study was carried out among the Bangladeshi healthcare workers from May 25, 2021, to June 5, 2021. The study found 422 responses from the healthcare workers of Bangladesh. Among the respondents, nearly half of them (45.26%) were doctors (n = 191). This study explored that the healthcare workers’ mucormycosis perception scores were significantly associated with their age, gender, profession, monthly income, marital status, job type, and death of friends and family members due to COVID-19. Conclusions This study emphasized the healthcare workers’ mucormycosis perception along with other associated factors. The findings could help policymakers to mitigate mucormycosis and related infectious diseases emergencies in the post-COVID-19...
Proceedings of the 3rd IEEE International Symposium on Signal Processing and Information Technology (IEEE Cat. No.03EX795)
Department of Electrical and Electronic Engineering, Bangladesh University of Engineering and Tec... more Department of Electrical and Electronic Engineering, Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh ... ABSTRACT This paper illustrates the superiority of using time-frequency varying averaging factor to estimate the a priori SNR far
Source camera model identification (CMI) and image manipulation detection are of paramount import... more Source camera model identification (CMI) and image manipulation detection are of paramount importance in image forensics. In this paper, we propose an L2-constrained Remnant Convolutional Neural Network (L2-constrained RemNet) for performing these two crucial tasks. The proposed network architecture consists of a dynamic preprocessor block and a classification block. An L2 loss is applied to the output of the preprocessor block, and categorical crossentropy loss is calculated based on the output of the classification block. The whole network is trained in an end-to-end manner by minimizing the total loss, which is a combination of the L2 loss and the categorical crossentropy loss. Aided by the L2 loss, the data-adaptive preprocessor learns to suppress the unnecessary image contents and assists the classification block in extracting robust image forensics features. We train and test the network on the Dresden database and achieve an overall accuracy of 98.15%, where all the test images are from devices and scenes not used during training to replicate practical applications. The network also outperforms other state-of-the-art CNNs even when the images are manipulated. Furthermore, we attain an overall accuracy of 99.68% in image manipulation detection, which implies that it can be used as a general-purpose network for image forensic tasks.
Ultrasound Shear Wave Elastography (USWE) with conventional B-mode imaging demonstrates better pe... more Ultrasound Shear Wave Elastography (USWE) with conventional B-mode imaging demonstrates better performance in lesion segmentation and classification problems. In this article, we propose SHEAR-net, an end-to-end deep neural network, to reconstruct USWE images from tracked tissue displacement data at different time instants induced by a single acoustic radiation force (ARF) with 100% or 50% of the energy in conventional use. The SHEAR-net consists of a localizer called the S-net to first localize the lesion location and then uses recurrent layers to extract temporal correlations from wave patterns using different time frames, and finally, with an estimator, it reconstructs the shear modulus image from the concatenated outputs of S-net and recurrent layers. The network is trained with 800 simulation and a limited number of CIRS tissue mimicking phantom data and is optimized using a multi-task learning loss function where the tasks are: inclusion localization and modulus estimation. Th...
Camera model identification has earned paramount importance in the field of image forensics with ... more Camera model identification has earned paramount importance in the field of image forensics with an upsurge of digitally altered images which are constantly being shared through websites, media, and social applications. But, the task of identification becomes quite challenging if metadata are absent from the image and/or if the image has been post-processed. In this paper, we present a DenseNet pipeline to solve the problem of identifying the source camera-model of an image. Our approach is to extract patches of 256*256 from a labeled image dataset and apply augmentations, i.e., Empirical Mode Decomposition (EMD). We use this extended dataset to train a Neural Network with the DenseNet-201 architecture. We concatenate the output features for 3 different sizes (64*64, 128*128, 256*256) and pass them to a secondary network to make the final prediction. This strategy proves to be very robust for identifying the source camera model, even when the original image is post-processed. Our mo...
Recognizing humor from a video utterance requires understanding the verbal and non-verbal compone... more Recognizing humor from a video utterance requires understanding the verbal and non-verbal components as well as incorporating the appropriate context and external knowledge. In this paper, we propose Humor Knowledge enriched Transformer (HKT) that can capture the gist of a multimodal humorous expression by integrating the preceding context and external knowledge. We incorporate humor centric external knowledge into the model by capturing the ambiguity and sentiment present in the language. We encode all the language, acoustic, vision, and humor centric features separately using Transformer based encoders, followed by a cross attention layer to exchange information among them. Our model achieves 77.36% and 79.41% accuracy in humorous punchline detection on UR-FUNNY and MUStaRD datasets – achieving a new state-of-the-art on both datasets with the margin of 4.93% and 2.94% respectively. Furthermore, we demonstrate that our model can capture interpretable, humorinducing patterns from al...
Removing speckle noise from medical ultrasound images while preserving image features without int... more Removing speckle noise from medical ultrasound images while preserving image features without introducing artifact and distortion is a major challenge in ultrasound image restoration. In this paper, we propose a multiframe-based adaptive despeckling (MADS) algorithm to reconstruct a high-resolution B-mode image from raw radio-frequency (RF) data that is based on a multiple input single output (MISO) model. As a prior step to despeckling, the speckle pattern in each frame is estimated using a novel multiframe-based adaptive approach for ultrasonic speckle noise estimation (MSNE) based on a single input multiple output (SIMO) modeling of consecutive deconvolved ultrasound image frames. The elegance of the proposed despeckling algorithm is that it addresses the despeckling problem by completely following the signal generation model unlike conventional ad-hoc smoothening or filtering based approaches, and therefore, it is likely to maximally preserve the image features. As deconvolution...
ESD is an important micro-parameter for ultrasonic tissue characterization. However, classifying ... more ESD is an important micro-parameter for ultrasonic tissue characterization. However, classifying between benign and malignant breast lesions from a broad dataset remains a challenge. In this work, we propose a new technique for the estimation of ESD of breast tissues from the diffuse component of backscattered RF data. This allows us to combine ESD with MSS and other ultrasonic macro-parameters for binary classification of lesions. In order to separate the diffuse component from the coherent component of the backscattered data, EEMD is performed. K-S test is used to automatically select the IMFs responsible for diffuse scattering. To ensure proper minimization of the system effects, a multi-step process is adopted where the RF data is deconvolved and filtered as the first two steps prior to normalization. As the ESD is supposed to have a fairly consistent value over a small tissue subregion, it is estimated, from the slope of the average regression line, computed from the exponentia...
An analysis of the noise effect on the convergence characteristic of the least-mean-squares (LMS)... more An analysis of the noise effect on the convergence characteristic of the least-mean-squares (LMS) type adaptive algorithms for blind channel identification is presented. It is shown that the adaptive blind algorithms misconverge in the presence of noise. A novel technique for ameliorating such misconvergence characteristic, using a frequency domain energy constraint in the adaptation rule, is proposed. Experimental results demonstrate that the robustness of the blind adaptive algorithms can be significantly improved using such constraints.
2002 11th European Signal Processing Conference, 2002
This paper presents an efficient thresholding technique for wavelet speech enhancement. The signa... more This paper presents an efficient thresholding technique for wavelet speech enhancement. The signal-bias compensated noise level is used as the threshold parameter. The noise as well as signal level is estimated from the detail wavelet packet (WP) coefficients in the first scale. Both hard and soft thresholding are applied successively. The regions for hard thresholding are identified by estimating their signal to noise ratio (SNR) in the wavelet domain. Soft thresholding is applied to the rest of the regions. The performance of the proposed scheme is evaluated on speech recorded in real conditions with artificial noise added to it.
Uploads
Papers by Kamrul Hasan