In the real world, a realistic setting for computer vision or multimedia recognition problems is that we have some classes containing lots of training data and many classes contain a small amount of training data. Therefore, how to use frequent classes to help learning rare classes for which it is harder to collect the training data is an open question. Learning with Shared Information is an emerging topic in machine learning, computer vision and multimedia analysis. There are different level of components that can be shared during concept modeling and machine learning stages, such as sharing generic object parts, sharing attributes, sharing transformations, sharing regularization parameters and sharing training examples, etc. Regarding the specific methods, multi-task learning, transfer learning and deep learning can be seen as using different strategies to share information. These learning with shared information methods are very effective in solving real-world large-scale problems. This special issue aims at gathering the recent advances in learning with shared information methods and their applications in computer vision and multimedia analysis. Both state-of-the-art works, as well as literature reviews, are welcome for submission. Papers addressing interesting real-world computer vision and multimedia applications are especially encouraged. Topics of interest include, but are not limited to:  • Multi-task learning or transfer learning for large-scale computer vision and multimedia analysis • Deep learning for large-scale computer vision and multimedia analysis • Multi-modal approach for large-scale computer vision and multimedia analysis • Different sharing strategies, e.g., sharing generic object parts, sharing attributes, sharing transformations, sharing regularization parameters and sharing training examples, • Real-world computer vision and multimedia applications based on learning with shared information, e.g., event detection, object recognition, object detection, action recognition, human head pose estimation, object tracking, location-based services, semantic indexing. • New datasets and metrics to evaluate the benefit of the proposed sharing ability for the specific computer vision or multimedia problem. • Survey papers regarding the topic of learning with shared information.  Authors who are unsure whether their planned submission is in scope may contact the guest editors prior to the submission deadline with an abstract, in order to receive feedback.

IEEE transactions on pattern analysis and machine intelligence

Paper-based immunoassays are becoming powerful and low-cost diagnostic tools, especially in resource-limited settings. Inexpensive methods for quantifying these assays have been shown using desktop scanners, which lack portability, and cameras, which suffer from the ever changing ambient light conditions. In this work, we introduce a novel approach of quantifying colors of colorimetric diagnostic assays with a smartphone that allows high accuracy measurements in a wide range of ambient conditions, making it a truly portable system. Instead of directly using the red, green, and blue (RGB) intensities of the color images taken by a smartphone camera, we use chromaticity values to construct calibration curves of analyte concentrations. We demonstrate the high accuracy of this approach in pH measurements with linear response ranges of 1–12. These results are comparable to those reported using a desktop scanner or silicon photodetectors. To make the approach adoptable under different lighting conditions, we developed a calibration technique to compensate for measurement errors due to variability in ambient light. This technique is applicable to a number of common light sources, such as sun light, fluorescent light, or smartphone LED light. Ultimately, the entire approach can be integrated in an “app” to enable one-click reading, making our smartphone based approach operable without any professional training or complex instrumentation.

Point-of-care colorimetric detection with a smartphone

A programmable logic integrated circuit device (“PLD”) includes programmable logic and a dedicated (i.e., at least partly hard-wired) digital signal processing region for performing or at least helping to perform digital signal processing tasks that are unduly inefficient to implement in the more general-purpose programmable logic and/or that, if implemented in the programmable logic, would operate unacceptably or at least undesirably slowly. The digital signal processing region may include multiple digital signal processing stages. The digital signal processing region may include a multiplier stage and one ore more stages that can operate in combination with the multiplier stage. The digital signal processing region has a plurality of modes such as for providing multiply-and-accumulate operation, multiply-and-add operation, etc.

Devices and methods with programmable logic and digital signal processing regions

Second generation high-performance PET scanners, called ClearPET$^{TM}$, have been developed by working groups of the Crystal Clear Collaboration (CCC). High sensitivity and high spatial resolution for the ClearPET camera is achieved by using a phoswich arrangement combining two different types of lutetium-based scintillator materials: LSO from CTI and LuYAP:Ce from the CCC (ISTC project). In a first ClearPET prototype, phoswich arrangements of 8$\times$8 crystals of 2$\times$2$\times$10 mm$^3$ are coupled to multi-channel photomultiplier tubes (Hamamatsu R7600). A unit of four PMTs arranged in-line represents one of 20 sectors of the ring design. The opening diameter of the ring is 120 mm, the axial detector length is 110 mm. The PMT pulses are digitized by free-running ADCs and digital data processing determines the gamma energy, the phoswich layer and even the exact pulse starting time, which is subsequently used for coincidence detection. The gantry allows rotation of the detector modules around the field of view. Preliminary data shows a correct identification of the crystal layer about (98$\pm$1)%. Typically the energy resolution is (23.3$\pm$0.5)% for the luyap layer and (15.4$\pm$0.4)% for the lso layer. Early studies showed the timing resolution of 2 ns FWHM and 4.8 ns FWTM. the intrinsic spatial resolution ranges from 1.37 mm to 1.61 mm full-width of half-maximum (FWHM) with a mean of 1.48 mm FWHM. further improvements in image and energy resolution are expected when the system geometry is fully modeled.

The ClearPET™ project: development of a 2nd generation high-performance small animal PET scanner

Since computer use is such a visually demanding task, vision problems and symptoms have become very common in today’s work place . Most studies indicate that computer operators who view their video display terminals (VDT) report more eye related problems than non VDT office workers. NIOSH Survey (National Institute of Occupational Safety and Health) has reported that visual symptoms occur in 75-90 % of VDT workers as opposed to 22 % musculoskeletal disorders (carpel tunnel syndrome) in computer users .

Computer vision syndrome.

A feasible way to gain the depth of interaction information in a positron emission tomography scanner is the use of phoswich detectors. In general, the layer of interaction is identified from the pulse shape of the corresponding scintillator material. In this work, pulses from LSO and LuYAP crystals were investigated in order to find a practical method of distinguishing. It turned out that such a pulse processing could be kept simple because of an additional slow component in the light decay of the LuYAP pulse. At the same time, the short decay time guarantees that the major amount of the light output is still collected within a short pulse recording time.

Pulse shape discrimination of LSO and LuYAP scintillators for depth of interaction detection in PET

An FPGA implementation guide for some different types of serial–parallel multiplier structures

Achieving robustness with a limited distortion level is a challenging design problem for watermarking systems in multimedia applications with a guaranteed quality requirement. In this paper, we provide an intelligent system for watermarking through incorporating a meta-heuristic technique along with an embedding method to achieve an optimized performance. The optimization objective is to provide the maximum possible robustness without exceeding a predetermined distortion limit. Hence, the quality level of the watermarking method could be guaranteed through that constraint optimization. A new fitness function is defined to provide the required convergence toward the optimum solution for the defined optimization problem. The fitness function is based on dividing its applied solution population into two groups, where each group is ranked according to a different objective. Thus, the multi-objectives in the problem are decoupled and solved through two single-objective sub-problems. Unlike existing watermarking optimization techniques, the proposed work does not require weighting factors. To illustrate the effectiveness of the proposed approach, we employ a recent watermarking technique, and then use it as the embedding method to be optimized. The Artificial Bee Colony is selected as the meta-heuristic optimization method in which the proposed fitness function is used. Experimental results show that the imposed quality constraint is satisfied, and that the proposed method provides enhanced robustness under different attacks for various quality thresholds. The presented approach offers a robust solution that can be applied to numerous multimedia applications such as film industry, intelligent surveillance and security systems.

A quality guaranteed robust image watermarking optimization with Artificial Bee Colony

The design of a robust watermarking technique has been always suffering from the conflict between the watermark robustness and the quality of the watermarked image. In this study, the embedding strength parameters for per-block image watermarking in the discrete cosine transform (DCT) domain are optimised. A fitness function is proposed to best suit the optimisation problem. The optimum solution is selected based on the quality and the robustness achieved using that solution. For a given image block, the peak-signal-to-noise ratio (PSNR) is used as a quality metric to measure the imperceptibility for the watermarked block. However, the robustness cannot be measured for a single watermark bit using traditional metrics. The proposed method uses the PSNR quality metric to indicate the degree of robustness. Hence, optimum embedding in terms of quality and robustness can be achieved. To demonstrate the effectiveness of the proposed approach, a recent watermarking technique is modified, and then used as the embedding method to be optimised. The Bees algorithm is selected as the optimisation method and the proposed fitness function is applied. Experimental results show that the proposed method provides enhanced imperceptibility and robustness under different attacks.

Quality metric-based fitness function for robust watermarking optimisation with Bees algorithm

The continuous development of forgery attacks on critical multimedia applications necessitates having accurate image tamper detection and localization techniques. In this paper, an image authentication approach is designed using a block-based fragile watermarking. Moreover, an effective recovery technique based on unsupervised machine learning is proposed. The authentication data is generated, for each 8 × 8 image block, using the Discrete Cosine Transform. A block dependency is established, for authenticating an image block, through using part of the authentication data of a distant block. Such block-dependency provides more accurate tamper detection and enables precise localization of tampered regions. At the recovery phase, a block is divided into smaller sub-blocks of size 2 × 2, where the recovery data is calculated through the K-means clustering. A fragile watermarking, in the spatial domain, is employed for embedding the watermark that is generated from the authentication and recovery data. We examine the effectiveness of the proposed approach under some of the most common attacks of image tampering, including copy move, constant average, and vector quantization attacks. Our approach is compared with several existing methods. Experimental results show that the proposed technique provides superior tampering detection and localization performance, and is capable of recovering the tampered regions more effectively.

H. I. Saleh

Papers

Pulse shape discrimination of LSO and LuYAP scintillators for depth of interaction detection in PET

An FPGA implementation guide for some different types of serial–parallel multiplier structures

A quality guaranteed robust image watermarking optimization with Artificial Bee Colony

Quality metric-based fitness function for robust watermarking optimisation with Bees algorithm

Fragile watermarking for image tamper detection and localization with effective recovery capability using K-means clustering