Combining Multiple Image Features to Guide Automatic Portrait Cropping for Rendering Different Aspect Ratios

Author(s):  
C S V C Cavalcanti ◽  
H M Gomes ◽  
J E R de Queiroz
2021 ◽  
Vol 104 (2) ◽  
pp. 003685042110113
Author(s):  
Xianghua Ma ◽  
Zhenkun Yang

Real-time object detection on mobile platforms is a crucial but challenging computer vision task. However, it is widely recognized that although the lightweight object detectors have a high detection speed, the detection accuracy is relatively low. In order to improve detecting accuracy, it is beneficial to extract complete multi-scale image features in visual cognitive tasks. Asymmetric convolutions have a useful quality, that is, they have different aspect ratios, which can be used to exact image features of objects, especially objects with multi-scale characteristics. In this paper, we exploit three different asymmetric convolutions in parallel and propose a new multi-scale asymmetric convolution unit, namely MAC block to enhance multi-scale representation ability of CNNs. In addition, MAC block can adaptively merge the features with different scales by allocating learnable weighted parameters to three different asymmetric convolution branches. The proposed MAC blocks can be inserted into the state-of-the-art backbone such as ResNet-50 to form a new multi-scale backbone network of object detectors. To evaluate the performance of MAC block, we conduct experiments on CIFAR-100, PASCAL VOC 2007, PASCAL VOC 2012 and MS COCO 2014 datasets. Experimental results show that the detection precision can be greatly improved while a fast detection speed is guaranteed as well.


Author(s):  
Kun Yang ◽  
Anning Pan ◽  
Yang Yang ◽  
Su Zhang ◽  
Sim Heng Ong

Remote sensing image registration with different viewpoints plays an important role in the field of geographic information system. However, when there exists ground relief variations and imaging viewpoint changes, non-rigid distortion occurs thus the registration becomes increasingly challenging. The current methods will suffer from missing true correspondences when non-rigid geometric distortion occurs. To address the problem, we propose a robust remote sensing image registration method based on SIFT feature distance and geometric structure features. At first, the scale-invariant feature transform (SIFT), a partial intensity invariant feature descriptor is used to extract reliable feature point set from sensed and reference image respectively. Secondly, a novel algorithm based on multiple image features which constrains the geometric structure during transformation is used to estimate exact correspondences between point sets. Finally, an accurate alignment is achieved by mapping the sensed image to reference image using thin-plate spline. We evaluated the performances of the proposed method by three sets of remote sensing images obtained from the unmanned aerial vehicle (UAV) and the Google earth, and compared with five state-of-the-art methods where our algorithm solved the non-rigid registration problem of remote sensing image with different viewpoints and showed the best alignments in most cases.


Author(s):  
Hiroto Nagayoshi ◽  
Takashi Watanabe ◽  
Tatsuhiko Kagehiro ◽  
Hisao Ogata ◽  
Tsukasa Yasue ◽  
...  

2008 ◽  
Vol 58 (3) ◽  
pp. 353-362 ◽  
Author(s):  
Jharna Majumdar ◽  
B. Vanathy ◽  
Lekshmi S.

Sign in / Sign up

Export Citation Format

Share Document