Imbalance Problems in Object Detection: A Review

By Kemal Oksuz, Baris Can Cam, Sinan Kalkan, Emre Akbas. 2019

568 citations recorded by the source corpus.

In this paper, we present a comprehensive review of the imbalance problems in object detection. To analyze the problems in a systematic manner, we introduce a problem-based taxonomy. Following this taxonomy, we discuss each problem in depth and present a unifying yet critical perspective on the solutions in the literature. In addition, we identify major open issues regarding the existing imbalance problems as well as imbalance problems that have not been discussed before. Moreover, in order to keep our review up to date, we provide an accompanying webpage which catalogs papers addressing imbalance problems, according to our problem-based taxonomy. Researchers can track newer studies on this webpage available at: https://github.com/kemaloksuz/ObjectDetectionImbalance.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

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

In Computer Vision and Pattern Recognition

1. Junwei Han
2. Gong Cheng
3. Hanwang Zhang
4. Dacheng Tao
5. Luc Van Gool
6. Bernt Schiele
7. Trevor Darrell
8. Xiwen Yao

Citations in the corpus

• Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentation (2020, neutral): “[42].”
• Detecting tiny objects in aerial images: A normalized Wasserstein distance and a new benchmark (2022, unknown): “The scale imbalance problem is another central issue hindering scale-invariant object detection (Zhang et al., 2017; Oksuz et al., 2021; Ming et al., 2019).”
• Improving Multispectral Pedestrian Detection by Addressing Modality Imbalance Problems (2020, background): “[33] present a comprehensive review of the imbalance problems in object detection and group these problems in a taxonomic tree with four main types: spatial imbalance, objective imbalance, class imbalance and scale imbalance.”
• YOLO-FaceV2: A Scale and Occlusion Aware Face Detector (2022, background): “During training, their cumulative contribution dominates the update of the model, leading to the overfit of the model [16].”
• Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, background): “Quite a number of surveys about object detection have been published in recent years [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], and our review differs from the existing ones mainly in two aspects.”
• REVISE: A Tool for Measuring and Mitigating Bias in Visual Datasets (2020, background): “Since then, researchers have noted the under-representation of object classes (Oksuz et al., 2019; Ouyang et al., 2016; Yang et al., 2014; Salakhutdinov et al., 2011; Buda et al., 2017; Liu et al., 2009), object contexts (Choi et al.”
• Unbiased Teacher for Semi-Supervised Object Detection (2021, background): “As mentioned in (Oksuz et al., 2020), in object detection, there exists foreground-background imbalance (e.g., background instances accounts for 70% of all training instances) and foreground classes imbalance (e.g., human instances accounts for 30% of all foreground training instances in MS-COCO…”
• Learning to Segment the Tail (2020, background): “[30] pointed out that the imbalance among different foreground categories, owing to the dataset itself, undermines the performance of popular recognition models.”
• IoU-uniform R-CNN: Breaking Through the Limitations of RPN (2019, background): “[24] review the deep-learning-era object detection literature and identify 8 different imbalance problems.”