New Benchmark for Household Garbage Image Recognition

Abstract

Household garbage images are usually faced with complex backgrounds, variable illuminations, diverse angles, and changeable shapes, which bring a great difficulty in garbage image classification. Due to the ability to discover problem-specific features, deep learning and especially convolutional neural networks (CNNs) have been successfully and widely used for image representation learning. However, available and stable household garbage datasets are insufficient, which seriously limits the development of research and application. Besides, the state-of-the-art in the field of garbage image classification is not entirely clear. To solve this problem, in this study, we built a new open benchmark dataset for household garbage image classification by simulating different lightings, backgrounds, angles, and shapes. This dataset is named 30 classes of household garbage images (HGI-30), which contains 18 000 images of 30 household garbage classes. The publicly available HGI-30 dataset allows researchers to develop accurate and robust methods for household garbage recognition. We also conducted experiments and performance analyses of the state-of-the-art deep CNN methods on HGI-30, which serves as baseline results on this benchmark.

Numerical Optimization Study of Combustion and Heat Transfer Characteristics in Municipal Solid Waste Incinerators

“…This is mainly because the height of the rear arch in case 4 is lower than that in other cases, and the flame is pressed toward the front wall by the lower rear arch. According to the distribution characteristics of the temperature field in the furnace under five working conditions, the height of the rear arch should not be too high or too low, and too high or too low may cause the high temperature corrosion of the water-cooled walls of the front and rear walls and other problems …”

Construct a garbage recognition model using automatic machine learning based on EasyDL

“…Plyukhin and Agha [13] provide a low-overhead delayed reference listing approach (DRL) for termination detection in actor systems as they examine the idea of autonomous garbage collection (GC) in the context of actor systems. By creating a new open benchmark dataset for household trash picture classification, dubbed 30 Classes of Household trash Images (HGI-30), Wu, et al [14] solve the issue of inadequate and unstable home garbage datasets. They also conduct experiments and performance analysis of state-of-the-art deep CNN methods on HGI-30.…”

Foreign object debris detection based on improved YOLOv5 algorithm

“…Therefore, compared with CIoU, EIoU divides the loss term of aspect ratio into the difference between the predicted width and height and the minimum outer frame width and height, which accelerates the convergence and improves the regression accuracy. Focal-EIoU Loss is introduced to reduce the optimization contribution of a large number of anchor boxes with less overlap with the target box to Bounding Box regression 10 , so that the regression process focuses on high-quality anchor boxes 11 .…”