Stereo correspondence using an assisted discrete cosine transform method

Author(s):  
Edward Rosales ◽  
Ling Guan
2021 ◽  
Author(s):  
Edward Rosales

Many approaches have been taken towards the development of a compliant stereo correspondence algorithm that is capable of producing accurate disparity maps within a short period of time. There has been great progress over the past decade due to the vast increase in optimization techniques. Currently, the most successful algorithms contain explicit assumptions of the real world such as definitive differences in disparity among objects and constant textures within objects. This thesis starts by giving a brief description of disparity, along with descriptions of some common applications. Next, it explores various methods used in common stereo correspondence algorithms, as well as gives an in depth description and analysis of top performing algorithms. These algorithms are later used to compare with the proposed algorithm. In the proposed algorithm, frequency stereo correspondence in parallel with the traditional color intensity stereo correspondence is used to develop an initial disparity map. Frequency stereo correspondence is achieved using a winner-take-all block based Discrete Cosine Transform (DCT) to find the largest frequency components as well as their positions to use in disparity estimation. The proposed algorithm uses methods that are computationally inexpensive to reduce the computational time that plagues many of the common stereo correspondence algorithms. The proposed algorithm achieves an average correct disparity rate of 95.3%. This results in a disparity error rate of 4.07% compared to the top performing algorithms in the Middlebury website [1]; the DoubleBP, CoopRegion, AdaptingBP, and ADCensus algorithms that have error rates of 4.19%, 4.41%, 4.23%, and 3.97%, respectively. Additionally, experimental results demonstrate that the proposed algorithm is computationally efficient and significantly reduces the processing time that plagues many of the common stereo correspondence algorithms.


2021 ◽  
Author(s):  
Nur Lukman ◽  
Jumadi Jumadi ◽  
Muhammad Faris Aminuddin ◽  
Dian Sa'adillah Maylawati ◽  
Nunik Destria Arianti ◽  
...  

2021 ◽  
Author(s):  
Edward Rosales

Many approaches have been taken towards the development of a compliant stereo correspondence algorithm that is capable of producing accurate disparity maps within a short period of time. There has been great progress over the past decade due to the vast increase in optimization techniques. Currently, the most successful algorithms contain explicit assumptions of the real world such as definitive differences in disparity among objects and constant textures within objects. This thesis starts by giving a brief description of disparity, along with descriptions of some common applications. Next, it explores various methods used in common stereo correspondence algorithms, as well as gives an in depth description and analysis of top performing algorithms. These algorithms are later used to compare with the proposed algorithm. In the proposed algorithm, frequency stereo correspondence in parallel with the traditional color intensity stereo correspondence is used to develop an initial disparity map. Frequency stereo correspondence is achieved using a winner-take-all block based Discrete Cosine Transform (DCT) to find the largest frequency components as well as their positions to use in disparity estimation. The proposed algorithm uses methods that are computationally inexpensive to reduce the computational time that plagues many of the common stereo correspondence algorithms. The proposed algorithm achieves an average correct disparity rate of 95.3%. This results in a disparity error rate of 4.07% compared to the top performing algorithms in the Middlebury website [1]; the DoubleBP, CoopRegion, AdaptingBP, and ADCensus algorithms that have error rates of 4.19%, 4.41%, 4.23%, and 3.97%, respectively. Additionally, experimental results demonstrate that the proposed algorithm is computationally efficient and significantly reduces the processing time that plagues many of the common stereo correspondence algorithms.


Author(s):  
Е.В. Костикова ◽  
Ш.С. Фахми ◽  
Я.А. Хасан ◽  
С.А. Селиверстов ◽  
А.В. Соболев ◽  
...  

Предложенный метод основан на использовании адаптивного косинусного преобразования в пространстве сигнала и по времени, и отличается произвольными размерами кубов в зависимости от пространственных и временных статистических характеристик сигнала изображений. Полученные результаты реализации предложенного метода в виде количества операции умножений и сложений двоичных данных. Предложен метод кодирования и декодирования изображений транспортной специфики, полученных из различных камер видеонаблюдения. Метод, основан на использовании адаптивного трехмерного дискретного косинусного преобразования, включающегося процедуры квантования, упорядочивания, адаптивного сканирования трансформант дискретного косинусного преобразования и энтропийного кодирования для передачи по каналу связи. Сжатие видеопоследовательностей, как правило, преследует две цели: уменьшить пространственную избыточность между пикселями изображения с учетом внутрикадровой корреляции и временную избыточность между последовательными кадрами с учетом межкадровой корреляции. Известные, на сегодняшний день все стандартны кодирования видео (H. 26-х и MPEG-х.) для снижения избыточности по времени используют принцип предсказания и компенсации движения по интерполированным позициям сэмплов в опорном кадре путём формирования векторов движения для предсказания кадров между опорными кадрами исходной видео последовательности. Реализация предложенного метода адаптивного косинусного преобразования в пространстве сигнала и по времени выполнена с использованием математического аппарата тригонометрических функций, позволяющих сократить общее число операции умножений и сложений, при этом рассмотренный метод отличается от известных стандартов произвольными размерами кубов в зависимости от пространственных и временных статистических характеристик сигнала изображений. Полученные результаты моделирования алгоритмов предложенного метода косинусного преобразования для размеров видео кубов (8х8,16х16, .512х512) в виде количества операций умножений и сложений двоичных данных по сравнению с алгоритмами построчно- столбикового преобразования и векторно-радикса. he proposed method is based on the use of adaptive cosine transform in the signal space and time, and differs in arbitrary cubic sizes depending on the spatial and temporal statistical characteristics of the image signal. The obtained results of the implementation of the proposed method in the form of the number of operations of multiplications and additions of binary data. A method for encoding and decoding transport-specific images obtained from various surveillance cameras is proposed. The method is based on the use of adaptive three-dimensional discrete cosine transform, including quantization, ordering, adaptive scanning transformants of the discrete cosine transform and entropy coding for transmission over the communication channel. Compression of video sequences, as a rule, has two objectives: to reduce spatial redundancy between image pixels with regard to intra-frame correlation and temporary redundancy between consecutive frames with regard to inter-frame correlation. Well-known, today all standard video coding (H. 26s and MPEG-x.) To reduce time redundancy use the principle of prediction and motion compensation for the interpolated positions of samples in the reference frame by forming motion vectors to predict frames between the reference frames of the original video sequences. Implementation of the proposed method of adaptive cosine transform in the signal space and time is performed using the mathematical apparatus of trigonometric functions that reduce the total number of operations of multiplications and additions, while the considered method differs from the known standards in arbitrary cubic sizes depending on the spatial and temporal statistical characteristics of the image signal . The obtained simulation results of the algorithms of the proposed cosine transform method for the size of video cubes (8х8.16х16, ..512х512) in the form of the number of operations of multiplications and additions of binary data in comparison with row-by-column transform and vector-radix algorithms.


Sign in / Sign up

Export Citation Format

Share Document