Theft Detection Algorithm Assignment

Uploaded by

veerpatel66930

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

29 views

Theft Detection Algorithm Assignment

Uploaded by

veerpatel66930

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 2

Theft Detection Algorithm Assignment

Company: Kudosware

Role: Data Scientist

This assignment is designed to assess your skills in building a theft detection algorithm using
video data, specifically focusing on shoplifting scenarios.

Dataset: DCSASS Dataset https://www.kaggle.com/datasets/mateohervas/dcsass-dataset

The DCSASS dataset contains a variety of video clips categorized into different activities,
including shoplifting. We will be focusing on videos labeled as "shoplifting" and "normal" to train
our theft detection model.

- Just use videos under shoplifting category

Task:

1. Data Preprocessing: Download the DCSASS dataset and filter for videos labeled as
"shoplifting" and "normal." Preprocess the videos by converting them into a suitable
format for your chosen machine learning model.
2. Feature Engineering: Extract relevant features from the preprocessed videos that can
be used to identify shoplifting activities.
3. Model Building: Develop a machine learning model to classify video frames/clips as
containing shoplifting or not containing shoplifting.
4. Evaluation: Evaluate the performance of your model using relevant metrics like:
○ Accuracy: Overall percentage of correct classifications.
○ Precision: Ratio of true positives (correctly identified shoplifting) to total positive
predictions.
○ Recall: Ratio of true positives to actual shoplifting instances in the dataset.
○ F1 Score: Harmonic mean of precision and recall, balancing both metrics.
5. Report: Prepare a report summarizing your approach, including:
○ Data preprocessing steps and rationale for chosen techniques.
○ Feature engineering methods and their effectiveness in identifying shoplifting.
○ Model selection with justification and hyperparameter tuning (if applicable).
○ Evaluation results, including confusion matrix and chosen metrics.
○ Discussion of strengths, weaknesses, and potential improvements to your
approach.

Deliverables:

● Jupyter Notebook or Python script containing your code

● Report summarizing your methodology, results, and discussion
Evaluation Criteria:

● Clarity and efficiency of code

● Appropriateness of feature engineering techniques
● Model selection and justification
● Performance of the model on the shoplifting detection task
● Communication and insights presented in the report

Time Limit:

Submit it by 26 April. Upload the required documents to google drive and make it public and
send it over on internshala.

Additional Notes:

● Feel free to explore libraries like OpenCV or TensorFlow for video processing and model
building.
● You can make any reasonable assumptions about shoplifting scenarios based on the
video descriptions or additional research.
● Be creative and showcase your problem-solving skills in building an effective shoplifting
detection algorithm using the DCSASS dataset.