Integral Image based Fast Convolution Operations via Separable Kernels

AbstractOne of the topical approaches in analysis – outside the framework of traditional ones – is the formation of an integral “image” of the object. There are several approaches to solving the issue of obtaining as much information about the sample by a certain portion of its properties or its composition as possible. The first approach is forming a visual image (diagram) of several different properties of the analyzed sample, for example, the content of certain metals, acids, volatile components and some other indicators of wine quality. The consolidated image of a sample enables us to distinguish samples identical or similar in the selected properties from crucially different ones, even in case of an acceptable change of each indicator. Or else, using the consolidated image one can evaluate the direction of an image shift of a certain sample compared to the set of standard samples. The analysis of the geometry of the sample image by diverse indicators affords ground for assumption of the reasons for this deviation, as well as identification of falsification, or even solution of a more complicated task: detecting the area of growth of raw materials. The second approach is close to the first one in terms of methodology, but it digitizes properties using detectors and presents this as an image (“visual print” of response) of signals of these detectors on some components of the sample (presence, content). The feature of this approach is the use of a detector system that is non-selective and cross-sensitive to certain sample components. These sample images are produced using a system of “electronic nose”. “Visual prints” of array signals of different character sensors contain qualitative and quantitative information about the part of the analyzed sample which is sorbed by sensors. Despite the uncertainty of this information, “electronic noses” of piezoelectric type are widely used in the analysis of samples with complex varying composition.

Download Full-text

Enabling Efficient Fast Convolution Algorithms on GPUs via MegaKernels

IEEE Transactions on Computers ◽

10.1109/tc.2020.2973144 ◽

2020 ◽

pp. 1-1

Author(s):

Liancheng Jia ◽

Yun Liang ◽

Xiuhong Li ◽

Liqiang Lu ◽

Shengen Yan

Keyword(s):

Fast Convolution

Download Full-text

Estimation of Moments on the Sphere by Means of Fast Convolution

Lecture Notes in Statistics - Statistical Challenges in Modern Astronomy V ◽

10.1007/978-1-4614-3520-4_47 ◽

2012 ◽

pp. 487-489

Author(s):

P. Bielewicz ◽

B. D. Wandelt ◽

A. J. Banday

Keyword(s):

Fast Convolution

Download Full-text

Fast convolution for recursive digital filters

IEEE Transactions on Audio and Electroacoustics ◽

10.1109/tau.1972.1162327 ◽

1972 ◽

Vol 20 (1) ◽

pp. 93-94 ◽

Cited By ~ 13

Author(s):

J. Meek ◽

A. Veletsos

Keyword(s):

Digital Filters ◽

Fast Convolution ◽

Recursive Digital Filters

Download Full-text

Analisis Pendeteksian Pola Wajah Menggunakan Metode Haar-Like Feature

JOURNAL OF INFORMATICS AND TELECOMMUNICATION ENGINEERING ◽

10.31289/jite.v2i2.2133 ◽

2019 ◽

Vol 2 (2) ◽

pp. 69 ◽

Cited By ~ 1

Author(s):

Sugandi Chau ◽

Jepri Banjarnahor ◽

Dikky Irfansyah ◽

Sinta Kumala

Keyword(s):

Edge Detection ◽

Bilateral Filter ◽

Cascade Classifier ◽

Integral Image ◽

Canny Edge Detection ◽

Canny Edge

Pencatatan kehadiran mahasiswa merupakan salah satu faktor penting dalam pengolahan kedisiplinan, kewajiban, dan ketaatan mahasiswa dalam mengikuti proses perkuliahan. Pencatatan kehadiran mahasiswa sebelumnya dilakukan dengan cara manual yaitu dengan menggunkan tanda tangan, pencatatan kehadiran dengan cara manual dapat menjadi penghambat pemantauan kedisiplinan, ketaatan mahasiswa dalam hal ketepatan waktu datang mahasiswa. Pencatatan kehadiran mahasiswa secara manual dapat diganti dengan pencatatan kehadiran mahasiswa secara terkomputerisasi yang menggunakan proses indentifikasi teknologi biometrik, untuk mengidentifikasi pola wajah mahasiswa digunakan metode bilateral filter, canny edge detection, haar-like feature, integral image, cascade classifier adaboost. Dari hasil pengujian, tingkat keberhasilan pencatatan kehadiran mahasiswa berdasarkan pola wajah dari masing-masing mahasiswa sebesar 70.43%.

Download Full-text

PENDETEKSIAN OBJEK ROKOK PADA VIDEO BERBASIS PENGOLAHAN CITRA DENGAN MENGGUNAKAN METODE HAAR CASCADE CLASSIFIER

International Journal of Natural Science and Engineering ◽

10.23887/ijnse.v1i3.12938 ◽

2018 ◽

Vol 1 (3) ◽

pp. 92

Author(s):

Kadek Oki Sanjaya ◽

Gede Indrawan ◽

Kadek Yota Ernanda Aryanto

Keyword(s):

Image Processing ◽

Detection System ◽

High Accuracy ◽

Special Study ◽

Cascade Classifier ◽

Integral Image ◽

Main Method ◽

Average Accuracy ◽

Haar Features ◽

Haar Cascade

Object detection is a topic widely studied by the scientists as a special study in image processing. Although applications of this topic have been implemented, but basically this technology is not yet mature, futher research is needed to developed to obtain the desired result. The aim of the present study is to detect cigarette objects on video by using the Viola Jones method (Haar Cascade Classifier). This method known to have speed and high accuracy because of combining some concept (Haar features, integral image, Adaboost, and Cascade Classifier) to be a main method to detect objects. In this research, detection testing of cigarettes object is in samples of video with the resolution 160x120 pixels, 320x240 pixels, 640x480 pixels under condition of on 1 cigarette object and condition 2 cigarettes object. The result of this research indicated that percentage of average accuracy highest 93.3% at condition 1 cigarette object and 86,7% in the condition 2 cigarette object that was detected on the video with resolution 640x480 pixels, while the percentage of accuracy lowest 90% at condition 1cigarette object, and 81,7% at the condition 2 cigarette objects, detected on the video with the lowest resolution 160x120 pixels. The percentage of average errors at detection cigarettes object was inversely with percentage of accuracy. So that the detection system is able to better recognize the object of the cigarette, then the number of samples in the database needs to be improved and able to represent various types of cigarettes under various conditions and can be added new parameters related to cigarette object

Download Full-text

Pendeteksian Bagian Tubuh Manusia untuk Filter Pornografi dengan Metode Viola-Jones

ComTech Computer Mathematics and Engineering Applications ◽

10.21512/comtech.v3i1.2447 ◽

2012 ◽

Vol 3 (1) ◽

pp. 482

Author(s):

Benny Senjaya ◽

Alexander A. S. Gunawan ◽

Jerry Pratama Hakim

Keyword(s):

Human Body ◽

Image Features ◽

The Internet ◽

Body Parts ◽

Cascade Classifier ◽

Integral Image ◽

Desktop Application ◽

Boosting Algorithm ◽

Similar Body ◽

Complete Framework

Information Technology does help people to get information promptly anytime and anywhere. Unfortunately, the information gathered from the Internet does not always come out positive. Some information can be destructive, such as porn images. To mitigate this problem, the study aims to create a desktop application that could detect parts of human body which can be expanded in the future to become an image filter application for pornography. The detection methodology in this study is Viola-Jones method which provides a complete framework for extracting and recognizing image features. A combination of Viola-Jones method with Haar-like features, integral image, boosting algorithm, and cascade classifier provide a robust detector for the application. First, several parts of the human body are chosen to be detected as the data training using the Viola-Jones method. Then, another set of images (similar body parts but different images) are run through the application to be recognized. The result shows 86.25% of successful detection. The failures are identified and show that the inputted data are completely different with the data training.

Download Full-text