# Image content analysis Location-aware mobile applications development Spring 2011 Paras Pant.

## Presentation on theme: "Image content analysis Location-aware mobile applications development Spring 2011 Paras Pant."— Presentation transcript:

Image content analysis Location-aware mobile applications development Spring 2011 Paras Pant

Overview Introduction Basic Image Analysis Content-Based Image Retrieval Some location based system.

Introduction Nowadays, the analysis of information has become paramount importance. Every image carries a huge amount of information but only a small part of it is relevant for a certain application. An image can be, grayscale or color, clear or foggy etc. Our interest is to analyze the content of the digital image.

Basics how many object. how many color object how many green.

Threshold Separate image object for the background. – O 1 = {f(m,n):f(m,n)>T} (object pixels) – B 2 = {f(m,n):f(m,n)<=T} (background pixels) – (m,n) value of pixel at some location, T threshold value Different thresholding algorithm exist. Global or local thresholding.

Gray Scale Image Thresholding Morphological operation imopen

Morphology operations on images The techniques used on these binary images. The foundation of morphological processing is in the mathematically rigorous field of set theory http://www.dspguide.com/ch25/4.htm

Image Threshold in every channel and combined Morphological operation imopen

Color Mean color given to the object Green color are given as feature and we extract the object

Clustering Grouping based on relevant information. K-mean Clustering 1.Define a cluster Centroid. 2.Assign each object to the group that has the closest centroid. 3.When all objects have been assigned, recalculate the positions of the K centroids. 4.Repeat Steps 2 and 3 until the centroids no longer move. This produces a separation of the objects into groups from which the metric to be minimized can be calculated. For more information refer to: http://cs.joensuu.fi/pages/franti/cluster/

Content- Based Image Retrieval (CBIR), Content-independent metadata: data that is not directly concerned with image content, but related to it. Examples are image format, author’s name, date, and location Content based metadata: such as color, texture and shape.

Content Based Image Retrieval Query Image Feature Calculation Similarity measure Preprocessing Database output Images pre processing Feature Calculation

Color based Feature and Similarity measure Histogram – Chromaticity histogram Preprocessing – Color space – Color Constancy » feature of the human color perception system which ensures that the perceived color of objects remains relatively constant under varying illumination conditions. Similarity measure – Distance Calculation – Cosine similarity – Histogram intersection

Transform image into different space CIE L* a* b* – Where L is Lightness, a & b color chromaticity HSV rg chromaticity r=R/R+G+B g=G/R+G+B

Image Retrieval Query Image Histogram Calculation Histogram Intersection Color Constancy Database output Images Color Constancy Histogram Calculation

Retrival results Query image

Result 2 Query image

Location Aware Application Image based location Awareness – http://137.189.32.220/welcome_files/Page391.ht m http://137.189.32.220/welcome_files/Page391.ht m – http://www.youtube.com/watch?v=V3V0UP4afUk http://www.youtube.com/watch?v=V3V0UP4afUk Image based Navigation.