Spot the Difference by Object Detection
This addresses the problem of efficient and accurate change detection for applications like verifying digital designs against photographic images, though it appears incremental as an adaptation of existing object detection methods.
The paper tackles the change detection task of spotting differences between two images by treating differences as objects to detect using a CNN-based object detection approach with an early-merging architecture. The method outperforms verification-based approaches by a large margin, provides location information, achieves 24 times acceleration through network compression while maintaining accuracy, and uses cheap synthetic training data without expensive bounding box annotations.
In this paper, we propose a simple yet effective solution to a change detection task that detects the difference between two images, which we call "spot the difference". Our approach uses CNN-based object detection by stacking two aligned images as input and considering the differences between the two images as objects to detect. An early-merging architecture is used as the backbone network. Our method is accurate, fast and robust while using very cheap annotation. We verify the proposed method on the task of change detection between the digital design and its photographic image of a book. Compared to verification based methods, our object detection based method outperforms other methods by a large margin and gives extra information of location. We compress the network and achieve 24 times acceleration while keeping the accuracy. Besides, as we synthesize the training data for detection using weakly labeled images, our method does not need expensive bounding box annotation.