SOPHON ACADEMY / Task
Introduction

Introduction

  • Objective

Develop a pet behavior recognition system based on the Milk-V Duo 256MB version that utilizes computer vision technology to analyze video data of pets, identify patterns in their behavior, and provide behavioral statistics reports. This will assist pet owners in better understanding their pets' activity habits, optimizing health management, and daily care for their pets.

 

  • Detailed Task Description

1. Development of Pet Behavior Recognition Algorithm
a. Develop a deep learning-based pet behavior recognition algorithm capable of identifying common behaviors from video streams, such as eating, playing, resting, and abnormal behavior.
b. The algorithm should be able to distinguish specific behavior patterns of different pet species.

 

2. Real-time Video Stream Processing and Analysis
a. Implement a video stream processing module for real-time capturing and analysis of video data input from cameras, extracting behavioral features of pets.
b. Ensure the latency of video stream processing is as low as possible for real-time updates of behavior recognition results.

 

3. Behavior Statistics and Reporting
a. Design a behavior statistics module to record the patterns and frequency of pet behavior over specific time periods.
b. Generate behavior statistics reports to help pet owners understand their pets' activity trends and potential health issues.

 

  • Performance Requirements
  1. The accuracy rate of the behavior recognition algorithm should be no less than 90%, with a false positive rate below 10%.
  2. The system should be capable of processing video streams of at least 720p resolution to ensure clarity in behavior recognition.
  3. Memory Usage: Optimize memory allocation while ensuring the efficiency of video processing and behavior recognition algorithms, to not exceed the 256MB memory limit.

 

  • Acceptance Criteria
  1. The pet behavior recognition algorithm runs stably on the Milk-V Duo and can accurately identify a variety of pet behaviors.
  2. The real-time video stream processing module can smoothly handle video streams, ensuring the timeliness of behavior recognition.
  3. The behavior statistics module can accurately record and display pet behavior data.
  4. The user interface is clear, easy to operate, and the setup process is intuitive.
  5. Functionality Testing: The system should pass at least 20 hours of pet behavior video testing to ensure stability and accuracy.
  6. The submitted project should include complete source code, documentation, and necessary resource files for subsequent maintenance and optimization.

 

By completing this task, developers will be able to demonstrate the application potential of the Milk-V Duo in the field of pet behavior analysis, providing a valuable tool for users to better care for their pets and enhance the quality of life for their pets.