Sampling Techniques for Large-Scale Object Detection from Sparsely Annotated Objects

Niitani, Yusuke; Akiba, Takuya; Kerola, Tommi; Ogawa, Teiichiro; Sano, Shotaro; Suzuki, Shuji

doi:10.48550/arxiv.1811.10862

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The weakly supervised object detection model aims to achieve better object detection performance in the case of sparse or inaccurate sample labels. [8] propose two methods for sample acquisition, using the partial perceptual sampling method to traverse the region candidates to obtain categories that are not related to the true labels, and discard the loss of these categories during the training process of the region candidates. [9] propose a method based on detection difference for weakly supervised semantic segmentation; [10] propose a progressive approach to train the weakly supervised semantic segmentation model to achieve layer-bylayer progressive optimization training.…”

Section: B Object Detection Based On Weakly Supervised Learningmentioning

confidence: 99%

A Self-Learning Object Detection Method Based on Highly Reliable Sample Mining

Li¹

2022

ijcsrr

View full text Add to dashboard Cite

Visual object detection is an artificial intelligence technique that locates specific objects from images, which is of great significance for practical applications. However, training general object detection models require many manually annotated images, bringing more labour and time cost. In order to improve the adaptability of the object detection model to the data environment changes, this paper proposes a self-learning object detection system based on high-reliability sample mining. We first train a SampleNet that can better mine reliable training samples from unlabeled data. We then use the combination of SampleNet and the basic object detection model to build a complementary residual training framework, continuously improving the sample mining ability and object detection tasks during the training process. The experimental results show that SampleNet can stably provide reliable pseudo samples for model training, and the complementary residual training framework improves the performance of basic object detection tasks.

show abstract

Section: B Object Detection Based On Weakly Supervised Learningmentioning

confidence: 99%

A Self-Learning Object Detection Method Based on Highly Reliable Sample Mining

Li¹

2022

ijcsrr

View full text Add to dashboard Cite

show abstract

Sampling Techniques for Large-Scale Object Detection from Sparsely Annotated Objects

Cited by 1 publication

References 0 publications

A Self-Learning Object Detection Method Based on Highly Reliable Sample Mining

A Self-Learning Object Detection Method Based on Highly Reliable Sample Mining

Contact Info

Product

Resources

About