Super-Resolution Network with Information Distillation and Multi-Scale Attention for Medical CT Image

The CT image is an important reference for clinical diagnosis. However, due to the external influence and equipment limitation in the imaging, the CT image often has problems such as blurring, a lack of detail and unclear edges, which affect the subsequent diagnosis. In order to obtain high-quality medical CT images, we propose an information distillation and multi-scale attention network (IDMAN) for medical CT image super-resolution reconstruction. In a deep residual network, instead of only adding the convolution layer repeatedly, we introduce information distillation to make full use of the feature information. In addition, in order to better capture information and focus on more important features, we use a multi-scale attention block with multiple branches, which can automatically generate weights to adjust the network. Through these improvements, our model effectively solves the problems of insufficient feature utilization and single attention source, improves the learning ability and expression ability, and thus can reconstruct the higher quality medical CT image. We conduct a series of experiments; the results show that our method outperforms the previous algorithms and has a better performance of medical CT image reconstruction in the objective evaluation and visual effect.

MGFPN: Enhancing multi-scale feature for object detection

Feature pyramids are commonly applied to solve the scale variation problem for object detection. One of the most representative works of feature pyramid is Feature Pyramid Network (FPN), which is simple and efficient. However, the fully power of multi-scale features might not be completely exploited in FPN due to its design defects. In this paper, we first analyze the structure problems of FPN which prevent the multi-scale feature from being fully exploited, then propose a new feature pyramid structure named Mixed Group FPN (MGFPN), to mitigate these design defects of FPN. Concretely, MGFPN strengthens the feature utilization by two modules named Mixed Group Convolution(MGConv) and Contextual Attention(CA). MGConv reduces the spatial information loss of FPN in feature generation stage. And CA narrows the semantic gaps between features of different receptive field before lateral summation. By replacing FPN with MGFPN in FCOS, our method can improve the performance of detectors in many major backbones by 0.7 to 1.2 Average Precision(AP) on MS-COCO benchmark without adding too much parameters and it is easy to be extended to other FPN-based models. The proposed MGFPN can serve as a simple and strong alternative for many other FPN based models.

Adaptive Weighted Multi-Level Fusion of Multi-Scale Features: A New Approach to Pedestrian Detection

Great achievements have been made in pedestrian detection through deep learning. For detectors based on deep learning, making better use of features has become the key to their detection effect. While current pedestrian detectors have made efforts in feature utilization to improve their detection performance, the feature utilization is still inadequate. To solve the problem of inadequate feature utilization, we proposed the Multi-Level Feature Fusion Module (MFFM) and its Multi-Scale Feature Fusion Unit (MFFU) sub-module, which connect feature maps of the same scale and different scales by using horizontal and vertical connections and shortcut structures. All of these connections are accompanied by weights that can be learned; thus, they can be used as adaptive multi-level and multi-scale feature fusion modules to fuse the best features. Then, we built a complete pedestrian detector, the Adaptive Feature Fusion Detector (AFFDet), which is an anchor-free one-stage pedestrian detector that can make full use of features for detection. As a result, compared with other methods, our method has better performance on the challenging Caltech Pedestrian Detection Benchmark (Caltech) and has quite competitive speed. It is the current state-of-the-art one-stage pedestrian detection method.

A Novel Bilinear Feature and Multi-Layer Fused Convolutional Neural Network for Tactile Shape Recognition

Convolutional neural networks (CNNs) can automatically learn features from pressure information, and some studies have applied CNNs for tactile shape recognition. However, the limited density of the sensor and its flexibility requirement lead the obtained tactile images to have a low-resolution and blurred. To address this issue, we propose a bilinear feature and multi-layer fused convolutional neural network (BMF-CNN). The bilinear calculation of the feature improves the feature extraction capability of the network. Meanwhile, the multi-layer fusion strategy exploits the complementarity of different layers to enhance the feature utilization efficiency. To validate the proposed method, a 26 class letter-shape tactile image dataset with complex edges was constructed. The BMF-CNN model achieved a 98.64% average accuracy of tactile shape. The results show that BMF-CNN can deal with tactile shapes more effectively than traditional CNN and artificial feature methods.