Weakly_detector
TensorFlow implementation of weakly-supervised object localization using only image-level labels, without bounding box annotations.
What is Weakly_detector?
Weakly_detector is a TensorFlow implementation of a weakly-supervised object localization method that can identify object locations in images using only image-level class labels. It implements the Class Activation Mapping (CAM) technique from the CVPR 2016 paper "Learning Deep Features for Discriminative Localization," which allows models to localize objects without requiring bounding box annotations during training. This solves the problem of expensive annotation costs in traditional object detection pipelines.
Computer vision researchers and practitioners working on object localization, especially those with limited annotated data or resources for bounding box labeling. It's also suitable for developers implementing weakly-supervised learning techniques in TensorFlow.
Developers choose Weakly_detector because it provides a clean TensorFlow implementation of a seminal weakly-supervised localization paper, enabling object detection with significantly reduced annotation requirements. The implementation is straightforward to adapt to custom datasets and integrates with standard pretrained models like VGG.
Overview
Tensorflow implementation of "Learning Deep Features for Discriminative Localization"
Use Cases
Best For
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
