Towards Unsupervised Crowd Counting via Regression-Detection Bi-knowledge Transfer

TL;DR

This paper proposes an unsupervised crowd counting method using bi-knowledge transfer between regression and detection models, enabling effective transfer learning from labeled source data to unlabeled target data.

Contribution

It introduces a novel mutual transformer-based knowledge distillation framework for unsupervised crowd counting via transfer learning.

Findings

Significant performance improvements on ShanghaiTech, UCF extunderscore CC extunderscore 50, and UCF extunderscore QNRF datasets.

Effective bi-knowledge transfer enables unsupervised crowd counting.

Outperforms existing state-of-the-art transfer learning methods.

Abstract

Unsupervised crowd counting is a challenging yet not largely explored task. In this paper, we explore it in a transfer learning setting where we learn to detect and count persons in an unlabeled target set by transferring bi-knowledge learnt from regression- and detection-based models in a labeled source set. The dual source knowledge of the two models is heterogeneous and complementary as they capture different modalities of the crowd distribution. We formulate the mutual transformations between the outputs of regression- and detection-based models as two scene-agnostic transformers which enable knowledge distillation between the two models. Given the regression- and detection-based models and their mutual transformers learnt in the source, we introduce an iterative self-supervised learning scheme with regression-detection bi-knowledge transfer in the target. Extensive experiments on standard crowd counting benchmarks, ShanghaiTech, UCF\_CC\_50, and UCF\_QNRF demonstrate a substantial improvement of our method over other state-of-the-arts in the transfer learning setting.