Self-Paced Learning Strategy with Easy Sample Prior Based on Confidence for the Flying Bird Object Detection Model Training

TL;DR

The paper targets Flying Bird Object Detection (FBOD) in surveillance video where difficulty varies across samples, and hard samples can degrade training. It introduces a Self-Paced Learning framework with Easy Sample Prior based on Confidence (SPL-ESP-BC), combining an ESP stage with a confidence-driven Minimizer to selectively weight samples in a one-class detection setting. The method yields an explicit Weighted Loss Function and a piecewise Minimizer, enabling progressive inclusion of harder samples as training progresses; it shows AP50 improvements of about 2.1% over standard training and achieves top performance among tested SPL variants. The proposed strategy robustly learns flying-bird characteristics from easy-to-hard contexts, reduces false detections, and demonstrates practical benefits for surveillance-based bird detection under noisy conditions.

Abstract

In order to avoid the impact of hard samples on the training process of the Flying Bird Object Detection model (FBOD model, in our previous work, we designed the FBOD model according to the characteristics of flying bird objects in surveillance video), the Self-Paced Learning strategy with Easy Sample Prior Based on Confidence (SPL-ESP-BC), a new model training strategy, is proposed. Firstly, the loss-based Minimizer Function in Self-Paced Learning (SPL) is improved, and the confidence-based Minimizer Function is proposed, which makes it more suitable for one-class object detection tasks. Secondly, to give the model the ability to judge easy and hard samples at the early stage of training by using the SPL strategy, an SPL strategy with Easy Sample Prior (ESP) is proposed. The FBOD model is trained using the standard training strategy with easy samples first, then the SPL strategy with all samples is used to train it. Combining the strategy of the ESP and the Minimizer Function based on confidence, the SPL-ESP-BC model training strategy is proposed. Using this strategy to train the FBOD model can make it to learn the characteristics of the flying bird object in the surveillance video better, from easy to hard. The experimental results show that compared with the standard training strategy that does not distinguish between easy and hard samples, the AP50 of the FBOD model trained by the SPL-ESP-BC is increased by 2.1%, and compared with other loss-based SPL strategies, the FBOD model trained with SPL-ESP-BC strategy has the best comprehensive detection performance.

Figures (6)

• Figure 1: Identifying the flying bird object in the surveillance video has different degrees of difficulty.
• Figure 2: The FBOD model training strategy based on SPL-ESP-BC.
• Figure 3: Illustration of calculating the prediction confidence for a bird object.
• Figure 4: The relationship between $\xi$ and training process.
• Figure 5: The detection effects of FBOD models are trained using four different training modes in three different situations.