Titlebook: Image and Graphics; 12th International C Huchuan Lu,Wanli Ouyang,Min Xu Conference proceedings 2023 The Editor(s) (if applicable) and The A

遵循的规范

反应

RatiO R-CNN: An Efficient and Accurate Detection Method for Oriented Object Detectionges, the oriented objects are distributed in any direction, and their ground truth bounding boxes have an extensive range of aspect ratios. In this paper, we propose a two-stage oriented object detection framework called RatiO R-CNN, which has good accuracy and efficiency on oriented object detectio

朴素

Student Classroom Behavior Detection Based on YOLOv7+BRA and Multi-model Fusioner, the current accuracy rate in behavior detection is low. To address this challenge, we propose the Student Classroom Behavior Detection system based on YOLOv7+BRA (YOLOv7 with Bi-level Routing Attention). We identified eight different behavior patterns, including standing, sitting, talking, liste

single

Tiny-YOLOv7: Tiny Object Detection Model for Drone Imagerye changes while flying lead to the scale of the object changes dramatically. In addition, drones flying quickly cause motion blur on the densely tiny objects. In order to address the two issues mention above, we propose Tiny-YOLOv7. In order to detect multi-scale objects, we replace the original pre

拒绝

Revisiting TENT for Test-Time Adaption Semantic Segmentation and Classification Head AdjustmentHowever, most test-time adaption methods made their success on classification tasks while object detection and segmentation tasks usually have more applications in the real world. Meanwhile, methods that update the model at test-time which is a main branch in test-time adaption (e.g., TENT [.], a ty

丰满有漂亮

Single Image Dehazing with Deep-Image-Prior Networkstmospheric scattering model to generate haze-free images, which may face the limitation of error accumulation. With the advance of deep learning technologies, employing deep neural networks (DNNs) to conduct haze removal becomes popular dehazing methods recently. Most DNNs-based methods automaticall

哑剧

conception

Dense Small Object Detection Based on Improved Deep Separable Convolution YOLOv5LOV5-G based on improved depth-wise separable convolution was proposed. The YOLOV5-G adds a prediction head based on the original YOLOv5 to improve the detection performance of small objects; The original loss function is changed to α-CIoU to obtain more accurate boundary box regression. The standar

确定方向

使苦恼

End-to-End Multilingual Text Recognition Based on Byte Modelingof each language cannot make full use of the information between different languages and consumes hardware resources very much, which makes the unified modeling of multiple languages very necessary. A natural approach to unified multilingual modeling is to combine modeling units (characters, subword