3D reconstruction from the microscopy images of serial sections plays an important role in analyzing structure of biological specimens, such as neuronal circuits in brain tissue. This paper is focusing on the 3D recon...
详细信息
ISBN:
(纸本)9781509023974
3D reconstruction from the microscopy images of serial sections plays an important role in analyzing structure of biological specimens, such as neuronal circuits in brain tissue. This paper is focusing on the 3D reconstruction of synapses which is a branch of micro reconstruction of brain. Having analysed the structure of synapses, we first detect and locate them in serial sections with cascade adaboost algorithm, then optimize the shape of synapses by constructing suitable fitting functions and segment them with morphology processing method. Taking the connection of adjacent sections and synaptic extensibility in space into consideration, we complete the reconstruction of synapses which is based on the segmentation in serial sections. Actually all our works are based on the registration of serial sections.
In this paper, a simple and fast algorithm was proposed to detect face. Firstly, some interest points marking skin regions were searched by only using simple chrominance Cr information instead of simultaneously using ...
详细信息
ISBN:
(纸本)9781479958368
In this paper, a simple and fast algorithm was proposed to detect face. Firstly, some interest points marking skin regions were searched by only using simple chrominance Cr information instead of simultaneously using chrominance Cr and Cb. And then, a conventional and popular adaboost algorithm was employed to make a decision whether around the detected interest points was there face and then located the face if there was. 400 face patches were translated into YCbCr space, and it was found from the 400 face patches that skin Cr histogram overlapped a little with non-skin Cr histogram and its cut-off was about 137. From LFW database, another 215 facial images of 108 subjects were random selected to test the proposed method. Compared with other interest detection methods, the detection based simple Cr method can not only find out all face regions, but also the number of its detected interest points was the least and the whole detection time was the least too. Therefore, the proposed method was an effective face detection method.
As an important branch of the computers,machine vision technology has a great influence on intelligent surveillance,human-computer interaction,and virtual *** paper introduced an intelligent emulator which used machin...
详细信息
ISBN:
(纸本)9781467377249
As an important branch of the computers,machine vision technology has a great influence on intelligent surveillance,human-computer interaction,and virtual *** paper introduced an intelligent emulator which used machine vision technology to guide *** intelligent emulator is controlled by the human *** hardware of intelligent emulator includes control unit,obstacle avoidance module,photosensitive module,DC motor module,*** sends instructions to microcontroller through the serial *** can control relay switch under the action of ULN2003AN when it received the instruction,then Microcontroller control the power modules which provide 6v dc to DC motor *** different levels make two DC motor rotate forward rotate or backward,and realize the emulator's *** the same time,the front,rear,left and right of the emulator were installed with the ultrasonic ranging module to avoid the *** the distance between the emulator and the obstacle is less than the safe distance of 40 cm,the emulator will *** addition,we set an appropriate threshold,the intelligent emulator will stop if the brightness is less than the *** resistance will determine whether the emulator meet the dark environment that the host computer cannot continue tracking the movement of head.
Stock index futures allows stock investors to manage different kinds of risk. This paper combines the adaboost feature selection and deep learning model for predicting stock index futures prices. In particular, a hybr...
详细信息
Stock index futures allows stock investors to manage different kinds of risk. This paper combines the adaboost feature selection and deep learning model for predicting stock index futures prices. In particular, a hybrid model is proposed in which the sklearn wrapped adaboost regressor is used for feature selection and the two-layer long short-term memory-based predictor is constructed. Performance metrics consistently show that the proposed model outperforms other popular prediction models such as random forest, multi-layer perception, gated recurrent unit, deep belief network and stacked denoising autoencoder.
A novel conception of automatic recognition in the way of no trouble is proposed for trouble of moving freight car detection system (TFDS), to solve the detection problem of sleeper springs in freight cars. The recogn...
详细信息
A novel conception of automatic recognition in the way of no trouble is proposed for trouble of moving freight car detection system (TFDS), to solve the detection problem of sleeper springs in freight cars. The recognition system inspires the feature pool of sleeper springs by Haar features, selects features by adaboost algorithm, builds a cascade of classifier, scans the whole image with detector of the classifier at every scale, and at last separates images with no trouble. It drastically reduces the amount of detected images and improves the manual recognition efficiency. Experiments show that the proposed method applies a set of simple features and an efficiency detecting strategy, performs high robustness against noise as well as transformation, rotation and scale of objects, and indicates high stability to the images with poor quality, such as low resolution, occlusion, poor illumination and excess exposure etc. The method can recognize sleeper springs in all-weather conditions, which advances the engineering application for TFDS.
The nucleus segmentation is the most important and tedious process in medical image analysis. The proposed method has three stages: preprocessing, h-maxima transformation based watershed segmentation and texture analy...
详细信息
ISBN:
(纸本)9781467393393
The nucleus segmentation is the most important and tedious process in medical image analysis. The proposed method has three stages: preprocessing, h-maxima transformation based watershed segmentation and texture analysis. First, the preprocessing stage uses top-hat filter to increase the contrast of nuclei and reduce the non-uniform illumination, imaging artifacts in the input image. In second stage, the segmentation of nuclei consists of a distance transformation, h-maxima transformation and watershed segmentation. The markers are used to obtain segments of the nuclei in the h-TMC watershed segmentation. To detect the single marker in nucleus, we usethese transformations. Due to imaging artifacts, prolonged cell cytoplasm in the contrast image, nuclei may falsely be segmented and it leads to an inaccurate analysis of the cell image. To identify and remove the non-nuclei segments. The third stage of texture analysis is followed. The texture with adaboost algorithm is used for non-nucleus identification.
Estimating head pose correctly from input face images is important for developing applications of vision-based human computer interaction. In this paper, we present a robust 3D head pose estimating approach using a sk...
详细信息
ISBN:
(纸本)9781467331487
Estimating head pose correctly from input face images is important for developing applications of vision-based human computer interaction. In this paper, we present a robust 3D head pose estimating approach using a skin-color model and haar-like feature detection. Using the detected primitive facial features of an obtained facial region, the corresponding 3D head pose of the 2D face is estimated. As for facial region detection, we adopt YC_bC_r model for skin-color model. For facial feature detection from the detected facial region, Haar-like feature is utilized along with adaboost learning. The 3D head pose of an input face image can be obtained by evaluating 3D information of facial features from the detected 2D eye-points and nose. Without any constraints for head pose estimation such as using initial face pose template of a frontal face, our proposed method can estimate face pose and retarget the motion to a 3D face model in real time. From the experiments, the proposed approach shows robustness in face and facial feature detection and eventually produces better results in estimating head pose rather than simply using Haar-like feature for both face and facial feature detection.
暂无评论