Cluster Centers Provide Good First Labels for Object Detection

Cluster Centers Provide Good First Labels for Object Detection

Burghouts, G.J.
Kruithof, M.C.
Huizinga, W.
Schutte, K.

2022

Learning object detection models with a few labels, is possible due to ingenious few-shot techniques, and due to clever selection of images to be labeled. Few-shot techniques work with as few as 1 to 10 randomized labels per object class. We are curious if performance of randomized label selection can be improved by selecting 1 to 10 labels per object class in a non-random manner. Several active learning techniques have been proposed to select object labels, but all started with a minimum of several tens of labels. We explore an effective and simple label selection strategy, for the case of 1 to 10 labels per object class. First, the full unlabeled dataset is clustered into N clusters, where N is the desired number of labels. Clustering is based on k-means on embedding vectors from a state-of-the-art pretrained image classification model (SimCLR v2). The image closest to the center is selected to be labeled. It is effective: on Pascal VOC we validate that it improves over randomized selection over 25%, with large improvements especially when having 1 label per object class. We have several benefits to report on this simple strategy: it is easy to implement, it is effective, and it is relevant in practice where one often starts with a dataset without any labels.

Clustering
Few labels
Label selection
Object detection
Computer vision
K-means clustering
Learning systems
Object recognition
Active Learning
Cluster centers
Clusterings
Label selection
Learning techniques
Object class
Performance
Simple++

http://resolver.tudelft.nl/uuid:0b4ca3f4-f811-482b-bcf8-aeff24443e4e

970944

Springer Science and Business Media Deutschland GmbH

9783031064265

0302-9743

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 21st International Conference on Image Analysis and Processing, ICIAP 2022, 23 May 2022 through 27 May 2022, 404-413

conference paper