A fast sea-level line extraction and object detection method for infrared sea image

Convolutional neural network (CNN)-based object detection is a key technology to enable autonomous mobile vision applications on mobile end devices such as smart phones and drones. With the advance of edge computing technology, a prevalent solution is to offload the computation-intensive CNN inference tasks to edge networks for fast and accurate object detection. However, a single edge server may not be powerful enough to ensure fast and accurate object detection due to its resource constraints. In this paper, we propose a multi-edge assisted fast object detection framework, MASS, to further reduce the object detection latency while maintaining the detection accuracy. In MASS, the CNN model is divided into two parts, namely the Head part and the Tail part. The Head part is executed locally and the Tail part is further split into multiple subtasks, which can be offloaded to multiple edge servers for parallel execution. First, we propose a method to select the optimal parallel entry point to separate the Head part from the Tail part. Then, an adaptive subtask generation and offloading strategy is proposed to divide the Tail part into multiple subtasks and offload these subtasks to multiple heterogeneous edge servers. Besides, we also propose a uniformly sampled zero-padding scheme to reduce the communication cost among edge servers when executing these subtasks in parallel. We implement MASS in a testbed with four edge servers and evaluate its performance, where the experimental results show that MASS can reduce object detection latency by up to $64.83%$, while the detection accuracy degradation is less than $3%$.

Shape template matching is an important approach in object detection and recognition. In this paper, we propose a fast and novel object detection method, which represents edge map contours with salient points and retrieves the target object by using a backtracking method with two stages from coarse matching to fine matching. Our method has two main contributions. One is we propose the way to represent contour structure with salient points. The other is that the object detection method proposed can directly operate on the real images, which improves its practicability. According to the experimental results, the proposed approach improves the speed than previous works.

With the development of science and technology, unmanned aerial vehicle (UAV) is more and more widely used to bring a lot of convenience to the society, but also led to serious threats to public security, personal privacy, military security and other aspects. Therefore, it is increasingly important to find unknown drones quickly and accurately. In UAV detection, the technologies based on acoustic, radio and radar detection are common, but these technologies usually require expensive equipment and strict configuration. However, the method based on machine vision has the advantages of low cost and simple configuration. In addition, detection and recognition methods based on deep learning have been fully developed, but most of them are for a single visible image, and the detection and recognition effect is limited. In this paper, a fast detection and identification method based is proposed based on the backbone of YOLOv3 (You Only Look Once version3). And dual-channel detectors were used as data sources. In this method, infrared and visible images are simultaneously input into the network for feature extraction, and the extracted depth features are concatenated. Then the multi-scale prediction network is used to regression the target location to obtain the final detection and recognition results. Finally, by collecting real UAV data sets, the network is trained and tested for comparative experiments. Experimental results show that the mAP of method in this paper is worthy of improvement, and the detection speed remains at 27images/s.

Fast and efficient object detection such as pedestrian detection and vehicle detection is an important task for autonomous driving. In camera-based sensing systems, object detection offers the fundamental ability for autonomous vehicles to perceive the dynamic environment in a real-time fashion. In this talk, I will be providing an overview of object detection based on computer vision and its applications for autonomous driving and the primary challenges as well as technical improvements in recent years. A number of topics will be covered in this talk, including the milestone detection approaches in history, detection datasets, performance metrics, speed-up techniques, and the recent state-of-the-art detection methods.

Fast near-duplicate keyframe detection is the basis of similar video content and video topic analysis. Traditional solutions based on local features are time-consuming and unsuitable for real-time applications. Considering there are little changes on the angles between similar video keyframes, a fast near-duplicate keyframe detection method based on local features is proposed. Firstly, the feature points are detected in various scales of image pyramids by adopting FAST (Features from Accelerated Segment Test) detection method. Secondly, each feature point is described by BRIEF algorithm. Thirdly, the similar keyframes are recognized based on the pattern entropy. Experiments prove that the proposed method is accurate and efficient, and is suitable for real-time applications of similar video keyframes detection.

The object detection algorithm based on vehicle-mounted lidar is a key component of the perception system on autonomous vehicles. It can provide high-precision and highly robust obstacle information for the safe driving of autonomous vehicles. However, most algorithms are often based on a large amount of point cloud data, which makes real-time detection difficult. To solve this problem, this paper proposes a 3D fast object detection method based on three main steps: First, the ground segmentation by discriminant image (GSDI) method is used to convert point cloud data into discriminant images for ground points segmentation, which avoids the direct computing of the point cloud data and improves the efficiency of ground points segmentation. Second, the image detector is used to generate the region of interest of the three-dimensional object, which effectively narrows the search range. Finally, the dynamic distance threshold clustering (DDTC) method is designed for different density of the point cloud data, which improves the detection effect of long-distance objects and avoids the over-segmentation phenomenon generated by the traditional algorithm. Experiments have showed that this algorithm can meet the real-time requirements of autonomous driving while maintaining high accuracy.

Forensic watermarking is often used to enable the tracing of digital pirates that leak copyright-protected videos. However, existing watermarking methods have a limited robustness and may be vulnerable to targeted attacks. Our previous work proposed a fallback detection method that uses secondary watermarks rather than the primary watermarks embedded by existing methods. However, the previously proposed fallback method is slow and requires access to all watermarked videos. This paper proposes to make the fallback watermark detection method faster using perceptual hashes instead of uncompressed secondary watermark signals. These perceptual hashes can be calculated prior to detection, such that the actual detection process is sped up with a factor of approximately 26,000 to 92,000. In this way, the proposed method tackles the main criticism about practical usability of the slow fallback method. The fast detection comes at the cost of a modest decrease in robustness, although the fast fallback detection method can still outperform the existing primary watermark method. In conclusion, the proposed method enables fast and more robust detection of watermarks that were embedded by existing watermarking methods.

Characterization of rice starch and protein obtained by a fast alkaline extraction method

International audience

AdaBoost algorithm is an efficient face detection method whose effectiveness is mainly influenced by the selection of weak classifier during the early process of training. To some extent, the selection of weak classifier depends on the selected sample set. Thus the training sample set is one of the most important factors in face detection. In this paper, the relationship between cascade classifier and weak classifier is analyzed in detail. Based on the factors of detection rate, undetected rate and false detection rate, an improved sample selection method is present and a fast face detection method which is divided into training and detection is proposed. The method is capable of optimizing the proportion of training samples and merging the detection window. The experimental results show that the proposed method is more effective than traditional ones.

Co-saliency detection, an emerging research area in saliency detection, aims to extract the common saliency from the multi images. The extracted co-saliency map has been utilized in various applications, such as in co-segmentation, co-recognition and so on. With the rapid development of image acquisition technology, the original digital images are becoming more and more clearly. The existing co-saliency detection methods processing these images need enormous computer memory along with high computational complexity. These limitations made it hard to satisfy the demand of real-time user interaction. This paper proposes a fast co-saliency detection method based on the image block partition and sparse feature extraction method (BSFCoS). Firstly, the images are divided into several uniform blocks, and the low-level features are extracted from Lab and RGB color spaces. In order to maintain the characteristics of the original images and reduce the number of feature points as well as possible, Truncated Power for sparse principal components method are employed to extract sparse features. Furthermore, K-Means method is adopted to cluster the extracted sparse features, and calculate the three salient feature weights. Finally, the co-saliency map was acquired from the feature fusion of the saliency map for single image and multi images. The proposed method has been tested and simulated on two benchmark datasets: Co-saliency Pairs and CMU Cornell iCoseg datasets. Compared with the existing co-saliency methods, BSFCoS has a significant running time improvement in multi images processing while ensuring detection results. Lastly, the co-segmentation method based on BSFCoS is also given and has a better co-segmentation performance.

Due to the sizes of synthetic aperture radar (SAR) images are very large, traditional deep learning methods usually require the dense sliding window pre-processing to obtain sub-images for ship detection. Generally, most sub-images contain no ship targets, and they may bring a large amount of redundant information that greatly affects the detection efficiency. To deal with the above problems, this paper proposes a fast ship detection method for SAR images in the inshore region, which mainly consists of two stages: scene classification and ship detection. The two stages employ two cascade deep convolutional networks, i.e., scene classification network (SCN) and single shot detector (SSD), respectively. The SCN can quickly eliminate the sub-images that may not contain ships, and then the remaining sub-images are input into the SSD to implement refined ship target detection. The experimental results based on the real dataset AIR-SARShip-1.0 show that the proposed method has higher efficiency than the original SSD method while maintaining relative high accuracy.

We propose a cascade method for fast and accurate traffic sign detection. The main feature of the method is that mid-level saliency test is used to efficiently and reliably eliminate background windows. Fast feature extraction is adopted in the subsequent stages for rejecting more negatives. Combining with neighbor scales awareness in window search, the proposed method runs at 3∼5 fps for high resolution (1360×800) images, 2∼7 times as fast as most state-of-the-art methods. Compared with them, the proposed method yields competitive performance on prohibitory signs while sacrifices performance moderately on danger and mandatory signs.

While accuracy provides quality, detection speed is paramount when it comes to usability. Several recently developed technologies have high accuracy eye detection; however, artificial intelligence technologies are moving away from fast detection. We focus on low computational loads and high speed for cases where eye detection is used. Therefore, a low brightness color threshold was applied to the existing projection function method, and an eyebrow removal operation was added. All of these operations were configured in parallel to operate in real time.

Non-maximum suppression (NMS) is an indispensable post-processing step in object detection. After a picture is detected by the target, a large number of redundant anchors will be obtained, so we need to go through NMS to process these repeated anchors. Its essence is to extract the target detection frame with high confidence and suppress the false detection frame with low confidence. The biggest problem in the NMS algorithm is that it completely removes adjacent low-confidence detection frames. In some images with dense targets, this is very likely to cause missed detection and false detection. Therefore, we proposed the Simple-NMS algorithm, which adds two new thresholds to the original NMS, and sets a constraint condition different from the original based on the new threshold, and does not change the complexity of the traditional NMS algorithm, which can be very good Improve the effectiveness of NMS missed targets. The new NMS algorithm is improved on the standard data sets PASAL VOC2007 and MS COCO. In addition, the algorithm is simple and efficient, and can be easily integrated into any other object detection process.