Faster ILOD: Incremental Learning for Object Detectors based on Faster RCNN

By Can Peng, Kun Zhao, Brian C. Lovell. 2020

148 citations recorded by the source corpus.

The human vision and perception system is inherently incremental where new knowledge is continually learned over time whilst existing knowledge is retained. On the other hand, deep learning networks are ill-equipped for incremental learning. When a well-trained network is adapted to new categories, its performance on the old categories will dramatically degrade. To address this problem, incremental learning methods have been explored to preserve the old knowledge of deep learning models. However, the state-of-the-art incremental object detector employs an external fixed region proposal method that increases overall computation time and reduces accuracy compared to object detectors such as Faster RCNN that use trainable Region Proposal Networks (RPNs). The purpose of this paper is to design an efficient end-to-end incremental object detector using knowledge distillation for object detectors based on RPNs. We first evaluate and analyze the performance of RPN-based detector with classic distillation towards incremental detection tasks. Then, we introduce multi-network adaptive distillation that properly retains knowledge from the old categories when fine-turning the model for new task. Experiments on the benchmark datasets, PASCAL VOC and COCO, demonstrate that the proposed incremental detector is more accurate as well as being 13 times faster than the baseline detector.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Meta & Continual Learning

Citations in the corpus

A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, methodology): “The idea of knowledge distillation is then introduced to other detection frameworks [283], [284], [285].”
EPicker is an exemplar-based continual learning approach for knowledge accumulation in cryoEM particle picking (2022, methodology): “Knowledge distillation 25 is a widely used technique for incremental object detection problems in natural images 26 – 28 , which transfers knowledge between neural networks using loss functions that minimize the difference between features extracted from the new and old models for old datasets.”
Energy-based Latent Aligner for Incremental Learning (2022, methodology): “For incremental object detection, similar to existing works [22, 37, 46], we follow a two task setting where the second task contains 10 , 5 or a single incremental class.”
Static-Dynamic Co-Teaching for Class-Incremental 3D Object Detection (2021, background): “Most of them (Shmelkov, Schmid, and Alahari 2017; Chen, Yu, and Chen 2019; Hao et al. 2019; Peng, Zhao, and Lovell 2020) address this problem by exploring knowledge distillation on network responses (i.e. data-based regularization).”
Towards Open World Object Detection (2021, methodology): “We use the standard protocol [61, 45] Table 3: We compare ORE against state-of-the-art incremental Object Detectors on three different settings.”
Incremental Object Detection via Meta-Learning (2020, neutral): “For the sake of comparison with [6], [24], we also report results on mini-val, which contains the first 5000 images from the validation split.”
Class-Incremental Learning: A Survey (2023, background): “Other works also apply knowledge distillation on the region proposal networks [263], [264], [265].”
OW-DETR: Open-world Detection Transformer (2021, methodology): “Different to standard object detection, incremental object detection approaches [27, 32] model newer object classes that are introduced in training incrementally and tackle the issue of catastrophic forgetting.”
Continual Object Detection via Prototypical Task Correlation Guided Gating Mechanism (2022, unknown): “For class-incremental object detection, the methods we compare include ILOD [45], Faster ILOD [39], ORE [23].”