Object Detection

What it does

This project demonstrates how to use TensorFlow Hub's pre-trained object detection models to identify and annotate objects within images. It covers:

Downloading and resizing images: Ensures images are in the right format and size for analysis. Running object detection: Utilizes TensorFlow Hub to detect objects and their locations. Drawing bounding boxes: Adds visual annotations to highlight detected objects.

How we built it

Image Handling: Employed Pillow for downloading, resizing, and processing images. Model Loading: Utilized TensorFlow Hub to load a pre-trained Faster R-CNN model for object detection. Bounding Box Visualization: Integrated Pillow to draw bounding boxes and labels on detected objects. Performance Measurement: Added time tracking to evaluate the efficiency of the detection process.

Challenges we ran into

Model Compatibility: Ensuring the TensorFlow Hub model URL was correct and compatible with the script. Image Format Issues: Handling different image formats and ensuring compatibility with TensorFlow's requirements. Dependency Management: Resolving conflicts and ensuring all necessary libraries were properly installed and updated.

Accomplishments that we're proud of

Seamlessly integrated TensorFlow Hub with image processing libraries to create a robust object detection pipeline.

What we learned

Model Handling: Gained valuable insights into loading and using pre-trained TensorFlow models from TensorFlow Hub. Image Processing: Enhanced understanding of image manipulation techniques for preparing data for machine learning models. Error Handling: Improved skills in troubleshooting and debugging issues related to model loading and image processing.

Built With

• python
• tensor-flow-hub
• tensorflow