Search

Purpose To compare the accuracy and repeatability of emerging machine learning based (i.e. deep) automatic segmentation algorithms with those of well-established semi-automatic (interactive) methods for determining liver volume in living liver transplant donors at computerized tomography (CT) imaging. Methods A total of 12 (6 semi-, 6 full-automatic) methods are evaluated. The semi-automatic segmentation algorithms are based on both traditional iterative models including watershed, fast marching, region growing, active contours and modern techniques including robust statistical segmenter and super-pixels. These methods entail some sort of interaction mechanism such as placing initialization seeds on images or determining a parameter range. The automatic methods are based on deep learning and they include three framework templates (DeepMedic, NiftyNet and U-Net) the first two of which are applied with default parameter sets and the last two involve adapted novel model designs. For 20 living donors (6 training and 12 test datasets), a group of imaging scientists and radiologists created ground truths by performing manual segmentations on contrast material-enhanced CT images. Each segmentation is evaluated using five metrics (i.e. volume overlap and relative volume errors, average/RMS/maximum symmetrical surface distances). The results are mapped to a scoring system and a final grade is calculated by taking their average. Accuracy and repeatability were evaluated using slice by slice comparisons and volumetric analysis. Diversity and complementarity are observed through heatmaps. Majority voting and Simultaneous Truth and Performance Level Estimation (STAPLE) algorithms are utilized to obtain the fusion of the individual results. Results The top four methods are determined to be automatic deep models having 79.63, 79.46 and 77.15 and 74.50 scores. Intra-user score is determined as 95.14. Overall, deep automatic segmentation outperformed interactive techniques on all metrics. The mean volume of liver of ground truth is found to be 1409.93 mL ± 271.28 mL, while it is calculated as 1342.21 mL ± 231.24 mL using automatic and 1201.26 mL ± 258.13 mL using interactive methods, showing higher accuracy and less variation on behalf of automatic methods. The qualitative analysis of segmentation results showed significant diversity and complementarity enabling the idea of using ensembles to obtain superior results. The fusion of automatic methods reached 83.87 with majority voting and 86.20 using STAPLE that are only slightly less than fusion of all methods that achieved 86.70 (majority voting) and 88.74 (STAPLE). Conclusion Use of the new deep learning based automatic segmentation algorithms substantially increases the accuracy and repeatability for segmentation and volumetric measurements of liver. Fusion of automatic methods based on ensemble approaches exhibits best results almost without any additional time cost due to potential parallel execution of multiple models.

In this paper, we present an approach that can detect scene nudity level with high precision using different deep net configurations. For this purpose, a recent approach [1] which has intense and very deep convolution layers is used. During net modelling, we strive to obtain most successful net configuration by comparing different Dropout models and image sizes -64 × 64, 128 × 128-. Additionally, leveraging the generalization capability of Support Vector Machine (SVM), improvement on success rate is demonstrated by retraining the features obtained at different output levels of the nets with SVM. At test and training stages, scene is investigated under three nudity levels: regular, semi-nudity and full-nudity. In order to evaluate false alarm rates of the net models, tests are conducted on different datasets which are SUN2012, LEAR Human and a dataset contains only semi-nudity samples besides validation set determined for each class. The results indicate that high precision rates can be achieved with low false alarm rate exploiting deep net models.

In this paper, we tackle content-based image-in-video search on large video archive. Particularly in this problem, the computation cost of a proposed method is another important parameter that should be considered along side of the success rate achieved due to the high volume of data. With our proposed method in this paper, content representation of multimedia data is achieved quite fast compared to other local approaches. Additionally, approximately % 16 improvement is introduced at the success rate with respect to the method that uses global approach for content representation on Stanford I2V dataset.

In this work, BoW, VLAD, and Hamming Embedding(HE) methods are examined that are used in image retrieval systems. Also shows that results can be improved using geometric consistency with binary signature obtained from Hamming Embedding. Oxford and Paris dataset are used for tests and obtained test results are comparatively examined over average precision rate. From results, observed that by adding geometric consistency information using binary signatures makes improvement on performance.

In this paper, we present a complete framework of an end-to-end error resilient transmission of 3D video over Digital Video Broadcasting - Handheld (DVB-H) and provide an extensive analysis of coding and transmission parameters. We perform the analysis for different coding and error resilience schemes using different contents coded at different bitrate levels. Throughout the experiments, we investigate the effects of video content type, video bitrate, coding method and unequal protection level for different channel conditions. The results show that Multi-view Coding (MVC) coding outperforms Simulcast and distribution of available bitrate between video quality and Forward Error Correction (FEC) protection is an important factor in different channel conditions.

In this paper, we propose a Multiple Description Coding (MDC) method for reliable transmission of compressed time consistent 3D dynamic meshes. It trades off reconstruction quality for error resilience to provide the best expected reconstruction of 3D mesh sequence at the decoder side. The method is based on partitioning the mesh vertices into two sets and encoding each set independently by a 3D dynamic mesh coder. The encoded independent bitstreams or so-called descriptions are transmitted independently. The 3D dynamic mesh coder is based on predictive coding with spatial and temporal layered decomposition. In addition, the proposed method allows for different redundancy allocations by duplicating a number of encoded spatial layers in both sets. The algorithm is evaluated with redundancy-rate-distortion curves and flexible trade-off between redundancy and side distortions can be achieved.

This paper presents an efficient joint source-channel coding scheme based on forward error correction (FEC) for three dimensional (3D) models. The system employs a wavelet based zero-tree 3D mesh coder based on progressive geometry compression (PGC). Reed-Solomon (RS) codes are applied to the embedded output bitstream to add resiliency to packet losses. Two-state Markovian channel model is employed to model packet losses. The proposed method applies approximately optimal and unequal FEC across packets. Therefore the scheme is scalable to varying network bandwidth and packet loss rates (PLR). In addition, distortion-rate (D-R) curve is modeled to decrease the computational complexity. Experimental results show that the proposed method achieves considerably better expected quality compared to previous packet-loss resilient schemes.

In this paper, a new face recognition system, GAYE, is presented. GAYE is a fully automatic system that detects and recognizes faces in cluttered scenes. The input of the system is any digitized image/image sequence that includes face/faces. The basic building blocks of the system are face detection, feature extraction and feature comparison. Face detection is based on skin color segmentation. For feature extraction, a novel approach is proposed that depends on the Gabor wavelet transform of the face image. By comparing facial feature vectors system finally makes a decision if the incoming person is recognized or not. Real time system tests show that GAYE achieves a recognition ratio over %90.