In order to accommodate multiple types of image encryption, a multi-image encryption algorithm for 2d and 3d images is designed in this paper. After recording the type and number of images, the pixels/coordinates of m...
详细信息
In order to accommodate multiple types of image encryption, a multi-image encryption algorithm for 2d and 3d images is designed in this paper. After recording the type and number of images, the pixels/coordinates of multiple images are stored in a cube block and are subjected to confusion anddiffusion operations. The confusion step uses the random length sequence position swapping method to swap a row (column) vector of variable length with another row (column) vector of the same length at a random position. The diffusion is done by Exclusive OR to combine pixels/coordinates at different locations with different chaotic matrices. Finally, the cipher images are output one by one. Experimental simulations and security analysis verify the effectiveness and security of the algorithm.
Medical image segmentation plays a vital role in disease diagnosis and analysis. However, data-dependent difficulties such as low image contrast, noisy background, and complicated objects of interest render the segmen...
详细信息
Medical image segmentation plays a vital role in disease diagnosis and analysis. However, data-dependent difficulties such as low image contrast, noisy background, and complicated objects of interest render the segmentation problem challenging. These difficulties diminish dense prediction and make it tough for known approaches to explore data-specific attributes for robust feature extraction. In this paper, we study medical image segmentation by focusing on robust data-specific feature extraction to achieve improveddense prediction. We propose a new deep convolutional neural network (CNN), which exploits specific attributes of input datasets to utilize deep supervision for enhanced feature extraction. In particular, we strategically locate anddeploy auxiliary supervision, by matching the object perceptive field (OPF) (which we define and compute) with the layer-wise effective receptive fields (LERF) of the network. This helps the model pay close attention to some distinct input data dependent features, which the network might otherwise 'ignore' during training. Further, to achieve better target localization and refineddense prediction, we propose the densely decoded networks (ddN), by selectively introducing additional network connections (the 'crutch' connections). Using five public datasets (two retinal vessel, melanoma, optic disc/cup, and spleen segmentation) and two in-house datasets (lymph node and fungus segmentation), we verify the effectiveness of our proposed approach in 2d and3d segmentation.
This paper describes a solution to automate the 3d model of an arbitrary object using an embedded UAV system through its movement and an image sensor to take pictures of an arbitrary static target, which is located wi...
详细信息
ISBN:
(数字)9781665497671
ISBN:
(纸本)9781665497671
This paper describes a solution to automate the 3d model of an arbitrary object using an embedded UAV system through its movement and an image sensor to take pictures of an arbitrary static target, which is located with a trajectory algorithm that allows the quadcopter to travel, photograph and come back. Then, the images are sent to a server, stored in a computer to run the photogrammetry process, and finally create the three-dimensional object without human intervention. It's important to mention that the algorithm parameters are set through an application that obtains the object's height, coordinates, size, and other parameters. Also, this kind of UAV device costs around200 dollars, and an application could control it, or if it's open-source, It could be configured by some programming language. This UAV system can be beneficial in different disciplinary fields, such as agriculture, construction, surveying, inventory control, surveillance, geology, biology, and other applications.
We propose an algorithm to segment 2d ellipses or 3d ellipsoids. This problem is of fundamental importance in various applications of cell biology. The algorithm consists of minimizing a contrast invariant energy defi...
详细信息
ISBN:
(纸本)9783319126104;9783319126098
We propose an algorithm to segment 2d ellipses or 3d ellipsoids. This problem is of fundamental importance in various applications of cell biology. The algorithm consists of minimizing a contrast invariant energy defined on sets of non overlapping ellipsoids. This highly non convex problem is solved by combining a stochastic approach based on marked point processes and a graph-cut algorithm that selects the best admissible configuration. In order to accelerate the computing times, we delineate fast algorithms to assess whether two ellispoids intersect or not and various heuristics to improve the convergence rate.
Our research focuses on developing an automated victim identification methodology for rescue robots in order to aid robot operators with the complex and stressful task of searching for victims in cluttered urban searc...
详细信息
Our research focuses on developing an automated victim identification methodology for rescue robots in order to aid robot operators with the complex and stressful task of searching for victims in cluttered urban search and rescue (USAR) environments. In this paper, we present an approach that utilizes 2d and3d sensory information from a real-time 3d sensory system for robust victim identification using both human geometric and skin region features. Our technique, uniquely, allows for the identification of partially occluded victims and single body parts that may be visible in cluttered USAR scenes using a Support Vector Machine-based classifier based on the aforementioned features. Unlike other approaches that focus on the recognition of one specific body part (such as the head) or the recognition of a small set of fixed body poses, we aim to identify multiple different body parts in a number of varying configurations to increase recognition rate. Experimental results illustrate the robustness of our methodology to find human victims in a variety of different poses in a rubble-filled USAR-like scene and its ability to potentially reduce operator workload.
Combinatorial maps describe the subdivision of objects in cells, and incidence and adjacency relations between cells, and they are widely used to model 2d and 3d images. However, there is no algorithm for comparing co...
详细信息
Combinatorial maps describe the subdivision of objects in cells, and incidence and adjacency relations between cells, and they are widely used to model 2d and 3d images. However, there is no algorithm for comparing combinatorial maps, which is an important issue for image processing and analysis. In this paper, we address two basic comparison problems, i.e., map isomorphism, which involves deciding if two maps are equivalent, and submap isomorphism, which involves deciding if a copy of a pattern map may be found in a target map. We formally define these two problems for nd open combinatorial maps, we give polynomial time algorithms for solving them, and we illustrate their interest and feasibility for searching patterns in 2d and 3d images, as any child would aim to do when he searches Wally in Martin Handford's books. (C) 2011 Elsevier Inc. All rights reserved.
It is known that regarding laser radar systems for highly precise 3d modelling the system performance is strongly influenced by the nonlinearities of the photoreceiver containing an optoelectronic device such as an MS...
详细信息
It is known that regarding laser radar systems for highly precise 3d modelling the system performance is strongly influenced by the nonlinearities of the photoreceiver containing an optoelectronic device such as an MSM-, PIN- or avalanche photodiode. In this paper we present nonlinear CAd models for photodiodes derived from measurements which can be used to analyse the nonlinear transfer function versus the optical large-signal stimulus. Accurate model correction of the distorted signals can be performed to make time interval measurements more accurate. This approach will be demonstrated on the basis of a MSM photodiode.
暂无评论