High efficient facial image compression is broadly required and challenging for surveillance and security scenarios, while either traditional general image codecs or special facial image compression schemes only heuri...
详细信息
ISBN:
(纸本)9781538644591;9781538644584
High efficient facial image compression is broadly required and challenging for surveillance and security scenarios, while either traditional general image codecs or special facial image compression schemes only heuristically refine codec separately according to face verification accuracy metric. We propose an End-to-End Facial Image Compression (E2EFIC) framework with a novel variable block size Regionally Adaptive Pooling (RAP) module whose parameters can be automatically optimized according to gradient feedback from an integrated semantic distortion metrics, including a successful exploration to apply Generative Adversarial Network (GAN) as metric directly in image compression scheme. The experimental results verify the framework's efficiency by demonstrating performance improvement of 71.41%, 48.28% and 52.67% bitrate saving separately over JPEG2000, WebP and neural network-based codecs under the same face verification accuracy distortion metric. We also evaluate E2EFIC's superior performance gain compared with latest specific facial image codecs.
Lossy compression of image and video yields visually annoying artifacts including blocking, blurring, ringing, etc., especially at low bit rates. In-loop filtering techniques can reduce these artifacts, improve qualit...
详细信息
ISBN:
(纸本)9781538644591;9781538644584
Lossy compression of image and video yields visually annoying artifacts including blocking, blurring, ringing, etc., especially at low bit rates. In-loop filtering techniques can reduce these artifacts, improve quality, and achieve coding gain accordingly. In this paper, we present a convolutional neural network (CNN) based in-loop filter for High Efficiency Video Coding (HEVC). First, we design a new CNN structure that is composed of multiple Variable-filter-size Residue-learning blocks, namely VRCNN-ext, for artifact reduction. VRCNN-ext is trained by natural images as well as their compressed versions at different quality levels. Second, we investigate a new in-loop filter based on the trained VRCNN-ext models. Specifically, we observed that using VRCNN-ext directly on the inter pictures is not effective. To solve this problem, we further train a classifier to decide whether to use VRCNN-ext for each coding unit (CU). The classifier makes decision based on the compressed information, thus avoiding the overhead bits to control the on/off of the CNN-based filter at the CU level. Experimental results show that our scheme achieves significant bits saving than the HEVC anchor, leading to on average 9.2%, 9.6% and 7.4% BD-rate reduction on the HEVC test sequences, under all-intra, low-delay B and random-access configurations, respectively.
The classification of large-scale high-resolution SAR land cover images acquired by satellites is a challenging task, facing several difficulties such as semantic annotation with expertise, changing data characteristi...
详细信息
Person re-identification is a crucial task of identifying pedestrians of interest across multiple surveillance camera views. For person re-identification, a pedestrian is usually represented with features extracted fr...
详细信息
Deep convolutional neural networks (DCNNs) have attracted much attention in remote sensing recently. Compared with the large-scale annotated dataset in natural images, the lack of labeled data in remote sensing become...
详细信息
—Existing generalization theories analyze the generalization performance mainly based on the model complexity and training process. The ignorance of the task properties, which results from the widely used IID assumpt...
详细信息
One key challenge to learning-based video compression is that motion predictive coding, a very effective tool for video compression, can hardly be trained into a neural network. In this paper we propose the concept of...
详细信息
In this paper, a novel framework for aircraft detection in high resolution apron area in Synthetic Aperture Radar (SAR) images is proposed, which combines the strength of location regression based convolutional neural...
详细信息
In this paper, a novel framework for aircraft detection in high resolution apron area in Synthetic Aperture Radar (SAR) images is proposed, which combines the strength of location regression based convolutional neural network (CNN) framework and the salient features of target in SAR images. Specifically, a Constant False Alarm Rate (CFAR) based target pre-locating algorithm is introduced, which can match the scale of target in SAR images more accurate compared to the existing region proposal method. In addition, in order to eliminate the fact of overfitting, we explore several strategies for SAR data augmentation, including translation, adding noise and rotation within a small range. Experiments are conducted on the data set acquired by the TerraSAR-X satellite in a resolution of 3.0 meters. The results show that the proposed detection framework could effectively obtain a more accurate detection result.
With the rapid development of urbanization, more and more attention has been paid to the structure of urban function zone. Thus, it is of great significance to investigate urban function zone. In this paper, we introd...
详细信息
ISBN:
(数字)9781728163741
ISBN:
(纸本)9781728163758
With the rapid development of urbanization, more and more attention has been paid to the structure of urban function zone. Thus, it is of great significance to investigate urban function zone. In this paper, we introduced the deep neural network (DNN) to infer the urban function zone with a supervised classification approach, taking the Shenzhen city in China as a case. First of all, the urban road networks of Shenzhen city were gathered and selected appropriately. Then, the fifth level road networks were utilized to segment the study region. Second, the communication data of different times and points of interest (POI) were collected. Then, the fifteen factors influencing urban function zone were derived. In addition, the urban function zone was divided into five types and the labeled examples with fifteen influencing factors were chosen. Third, the labeled examples were employed to train the DNN with different hidden layers compared with random forest (RF) and support vector machine (SVM). The models were trained with the approach of five-fold cross validation, and the average training accuracy with five times is taken as the accuracy of models. Finally, this paper compared the accuracy. It's been shown in the results that DNN was the optimum model and achieved the highest accuracy. Therefore, our proposed method is an efficient approach to infer the urban function zone.
Synthetic aperture radar (SAR) allows all-weather, day and night surveillance. Thus, it is of great significance for the ship detection and recognition. Because of the SAR special imaging mechanism, it is very difficu...
详细信息
ISBN:
(数字)9781728163741
ISBN:
(纸本)9781728163758
Synthetic aperture radar (SAR) allows all-weather, day and night surveillance. Thus, it is of great significance for the ship detection and recognition. Because of the SAR special imaging mechanism, it is very difficult to extract the ship features with SAR image for the traditional target detection algorithm. In this paper, we proposed a approach which is composed of you only look once (YOLO) algorithm, sliding window detection strategy, and clustering algorithm. Firstly, the SAR images of GaoFen-3 and training dataset are gathered. Secondly, the experiments about the size of ship detection frame is carried out to find the optimum size of the frame for the training model. Thirdly, the ships are detected initially with YOLO v3 and fast region-based convolutional neural network (Fast-RCNN). Finally, the detected ships are clustered adaptively, and the experimental results of YOLO v3 and Fast-RCNN are compared and discussed at length. Our experimental results demonstrated that our method outperformed Fast-RCNN to detect the ships in the surface sea with low-resolution wide -band SAR images. Therefore, our approach is a robust method to detect the ships in the surface sea with SAR images.
暂无评论