MEDIA SIGNAL PROC
MEDIA SIGNAL PROC MAT 201A
Popular in Course
Popular in Media Arts and Design
This 30 page Class Notes was uploaded by Marguerite Wintheiser on Thursday October 22, 2015. The Class Notes belongs to MAT 201A at University of California Santa Barbara taught by Staff in Fall. Since its upload, it has received 52 views. For similar materials see /class/227002/mat-201a-university-of-california-santa-barbara in Media Arts and Design at University of California Santa Barbara.
Reviews for MEDIA SIGNAL PROC
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 10/22/15
UCSB If Phillip J Popp and Jamie Timms EMGbased input 201A Final Presentation Introduction to EMGbased input BCI usage scenario for alternative form of computer mediated communication and interaction Explore various types of EMG signal processing techniques in order to recognize meaningful information in physiological data Three stages of pre processing processing and post processing EMG signal Problems and solutions EMG Signal EMG stands for electromyography It is the study of muscle electrical signals This signal is normally a function of time and is describable in terms of its amplitude frequency and phase Simple model ofthe EMG signal rl 397l ZM ldlz 7 I ll w1 1 yin User Scenario ALTERNATIVE BRAINCOMPUTER INTERFACE BCI eigsytasy USAGE SCENARIO serenin tenov annoyance ve I trust Mapping emotional states to i loathing muscle activities admiration EMOISE contempt 39emmse amazemem boredom disgust loathing 39 5 Sam s disapproval griei sadness pensiveness System Architecture 39 EIECtrOde Placement 8bit 8khz MuLaw mono wav 16 bit 16 khz PCM mono aif File Transfer MATLAB Processing A MATLAB FTP EMG MP4O gt iPhone PWP Problems External Noise Internal Noise Signal Distortion from Denoising Process Classi cation of EMG Signals U39lbUJNH Denoise Segment Process Extract Features Analyze Solutions Segment Find Zero Windowed RMS Low Pass Filter Median Filter Differentiate Crossing Peaks and Troughs A M N l f 391 k e KVJ Rik W U Process Complex Wavelet Transform Nearly Shift Invariant as opposed to the Discrete Wavelet Transform Robust Against Denoising Techniques a Duaerree Complex Tree 1 7 Wavelet Transform gum Q Structure 39 a xn 4 gt 39 439 Ireeh a Experiment Iteration Process Ten Repetition Muscle Activity A Magnitude of Level 1 Approximation Coef cients of Kingsbury DualTree Complex Wavelet Transform Wavelet Coefficient Experiment Iteration a m m N Ten Repetition of Muscle Activity B Magnitude of Level 1 Approximation Coefficients of Kingsbury Dual Tree Complex Wavelet Transform Extract Features Two Datasets 1 Down Sampled Complex Wavelet Coef cients 2 Signal to Noise Ratio Full Wave complex Low Pass Wavelet Down Sample Rectification Filter Transform Advanced SNR Segmentation Estimation Reasoning Strong Correlation Between Complex Wavelet Coef cients andSNR SNR Acts as Con dence Factor During Analysis Stage Analyze Hidden Input Neural Network Output Supervised Training Large Training Data Set Vision Man Vs Machine By Jason Pele Vision Taken for Granted 0 Much about human vision remains unclear o How does the brain work with the eye to form perceptions and interpretations of what we see 0 Billions of neural cell connections within the brain present a large complexity problem 0 Yet human vision is fallible Illusions and ambiguities are encountered all the time Computer Vision 0 Development of a theoretical and algorithmic base for which useful information about the world can be extracted and analyzed from images 0 Goal it to comprise a computer system that is closely modeled after the human visual system 0 Image formation low and high level processing and then 3D descriptioninterpretation Developing CV based on HVS 3D scene description Low Level Processing What a computer sees and ways to analyze it Matrix of pixels where values usually represent the intensity of image at that point 0 In MATLAB there are different image types usually grayscale and RGB are used where RGB is a NxNx3 array 0 Simple feature computations intensity color edges 0 Organization grouping of pixels regions lines o Higherfeature computations patterns textures geometrical shape descriptions RGB image GREG and histogram mensny prom a ong Greg s sca es Intensity profile tenswy va ues so 80 100 120 140 160 1 Dmance a ong prome Prome a ong ground p asuc contamer 500 600 7 y 0 100 200 RGB image Samurai and histogram Samuvaw an a a mane Number of P Xe s 7000 6000 5000 4000 3000 2000 1000 Samuraw sttogram 100 150 ntenswty or RGB va ues v 200 x 250 Samuvaw an a M mane mtensny vames Samuraw 2mm sun ADD Dwstance a ung pmme sun Edge detection o Using the Sobel method Spatial gradient measurement using two 3x3 convolutional kernels Perceptions or Hypotheses o How much of what we perceive is based on prior knowlede or exerience Viewing the 3D World 0 We can get a lot of information from a 2D image but what about distances of objects curvatures shapes etc o Humans are equipped with binocular vision two eyes 0 Computer vision must then utilize multiple images at multiple angles from multiple calibrated cameras Homography 2D to 2D mapping 0 Maps points in an image plane as seen from one camera to points in the same image plane as seen from a different camera 0 Useful for determining positioning of objects in a picture 0 Introduces the problem in CV known as point correspondence that becomes even more difficult in a 3D to 2D mapping Samuvaw at an ang e mapped m M mane MATH a aways MAT a an ang e Summary CV is difficult because we don t have a concrete model HVS to base it on 0 We must take an array of numbers as an input and come up with ways to describe textures colors depth objects and even more complex interpretations such as action and situation a The computer can see the input imagescene in many different ways the key is to understand how to combine these representations to reach the goal of having a machine or computer view the world as a human does Other relative topics Biological vision Image and signal processing Neurophysiology Psychology Psychophysics Other Cognitive sciences o The list goes on
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'