A computer vision project that performs object detection on images using Python and OpenCV's Deep Neural Network (DNN) module. The system utilizes a pre-trained SSD MobileNet V3 model trained on the COCO dataset to accurately identify and label multiple objects within a scene.
- Real-time Object Detection: Detects and classifies objects using a pre-trained MobileNet model.
- COCO Dataset Classes: Capable of recognizing up to 91 different everyday object categories (people, vehicles, animals, furniture, etc.).
- Visual Annotations: Automatically draws bounding boxes around detected objects and displays the predicted class name.
- Easily Adaptable: Configured to run on static images (like
lena.jpg), but can be easily uncommented to run on a live webcam feed.
- Language: Python 3
- Computer Vision: OpenCV (
cv2) - Model: SSD MobileNet V3 Large (COCO 2020_01_14)
main.py: The main script that loads the model, processes the image, and displays the output.coco.names: Text file containing the list of class labels the model can predict.ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt: The configuration file for the neural network architecture.frozen_inference_graph.pb: The frozen weights of the pre-trained TensorFlow model.lena.jpg/lena.png: Sample image used for testing the detection script.
- Make sure you have Python installed on your system.
- Install the required OpenCV package:
pip install opencv-python