The Effects of Mandarin Chinese Lexical Tones in Sound–Shape and Sound–Size Correspondences

Crossmodal correspondences refer to when specific domains of features in different sensory modalities are mapped. We investigated how vowels and lexical tones drive sound–shape (rounded or angular) and sound–size (large or small) mappings among native Mandarin Chinese speakers. We used three vowels (/i/, /u/, and /a/), and each vowel was articulated in four lexical tones. In the sound–shape matching, the tendency to match the rounded shape was decreased in the following order: /u/, /i/, and /a/. Tone 2 was more likely to be matched to the rounded pattern, whereas Tone 4 was more likely to be matched to the angular pattern. In the sound–size matching, /a/ was matched to the larger object more than /u/ and /i/, and Tone 2 and Tone 4 correspond to the large–small contrast. The results demonstrated that both vowels and tones play prominent roles in crossmodal correspondences, and sound–shape and sound–size mappings are heterogeneous phenomena.

Subspace Detours Meet Gromov–Wasserstein

In the context of optimal transport (OT) methods, the subspace detour approach was recently proposed by Muzellec and Cuturi. It consists of first finding an optimal plan between the measures projected on a wisely chosen subspace and then completing it in a nearly optimal transport plan on the whole space. The contribution of this paper is to extend this category of methods to the Gromov–Wasserstein problem, which is a particular type of OT distance involving the specific geometry of each distribution. After deriving the associated formalism and properties, we give an experimental illustration on a shape matching problem. We also discuss a specific cost for which we can show connections with the Knothe–Rosenblatt rearrangement.

Multimedia Concepts on Object Detection and Recognition with F1 Car Simulation Using Convolutional Layers

This paper presents a framework for detecting objects in images based on global features and contours. The first step is a shape matching algorithm that uses the background subtraction process. Object detection is accomplished by an examination of the oversegmentation of the image, where the space of the potential boundary of the object is examined to identify boundaries that have a direct resemblance to the prototype of the object type to be detected. Our analysis method removes edges using bilinear interpolation and reestablishes color sensors as lines and retracts background lines from the previous frame. Object contours are generated with clustered lines. The objects detected will then be recognized using the extraction technique. Here, we analyze the color and shape characteristics with which each object is capable of managing occlusion and interference. As an extension of object detection and recognition, F1 car simulation is experimented with simulation using various layers, such as layer drops, convolutionary layers, and boundary elimination, avoiding obstacles in different pathways.

Real-Time Detection of Linear Structure Objects Using Mean Shift Segmentation and Heuristic Search

With the development of intelligent industrial production, industrial components with linear structure tend to be regular, such as TV LCD module, mobile phone screen, and electronic equipment shell. Recognition of linear structure objects by machine vision is an important aspect of intelligent industry. At present, shape matching algorithm is mostly used for arbitrary structure objects. It will be time-consuming if it is directly used to detect the linear structure objects as it needs to traverse the parameter space of the object. To solve the traversal problem and detect the linear structure objects in real time, a heuristic detection algorithm is designed according to the characteristics of linear structure objects. First, the coarse position and orientation are obtained by mean shift filtering and heuristic region grouping to reduce the searching range. Then, the heuristic search method is used to get the precise location information. The heuristic search method is designed based on the particle swarm optimization algorithm and heuristic information. The proposed method has been evaluated on two image databases of common industrial parts and backlight units which are typical linear structure objects. The experimental results showed that the proposed algorithm could reduce the detect time by more than 70% averagely while the detection accuracy is kept. It proves that the proposed algorithm can detect linear structure objects in real time and is suitable for the detection of objects with linear structures.

Virtual Interface with Kinect 3D Sensor for Interaction with Bedridden People

The human-machine interaction has evolved significantly in the last years, allowing a new range of opportunities for developing solutions for people with physical limitations. Natural user interfaces (NUI) allow bedridden and/or physically disabled people to perform a set of actions trough gestures thus increasing their quality of life and autonomy. This paper presents a solution based on image processing and computer vision using the Kinect 3D sensor for development of applications that recognize gestures made by the human hand. The gestures are then identified by a software application that triggers a set of actions of upmost importance for the bedridden person, for example, trigger the emergency, switch on/off the TV or control the bed slope. It was used a shape matching technique for six gestures recognition, being the final actions activated by the Arduino platform. The results show a success rate of 96%. This system can improve the quality of life and autonomy of bedridden people, being able to be adapted for the specific necessities of an individual subject.