Sunday, April 20, 2014


Topic: "Data Analysis"
Client: Mr. Ramahi

Short Descp:
There is data consist of twelve sensor measurement result that need to be clustered. These data comes from twelve sensor attached on a hand of 19 people. They move their hand in 11 managed style. The clustering method used in this observation is unsupervised SOM.

Image Reconstruction from Noisy Image

Topic: "Digital Image Processing"
Client: CarrierGuide15

Short Descp:
In this project, I was trying to reconstruct image from noisy one without losing much detail information in it. In Gonzalez amazing book, they provide a method named Backprojection to reconstruct image. What I did is to implement this method and design new scheme to raise result. In this project, I uses Radon + Hamming as an improvement scheme.

Face Recognition in Various Dark Light and Dim light Conditions

Topic: "digital image processing"
Client: Mr. Saurabh Parmar

Short Descp:
What I dealt on this project was to design quite good system to recognize faces under dark and dim light condition. Until now, light condition around objects (faces) to recognize is still an open problem. Many scientist is still observing better way to resolve it. This is a master thesis project. Accuracy and recognition speed are main issue to consider in the development. Improvement has to be developed based on the two base papers. So, the idea is not too choose or design new method for feature extraction and recognition, but more to design scheme for pre-processing steps. When we are dealing

Predicting CPU Performance Using Learning Machine Compared to Linear Regression Method

Topic: "Machine Learning"
Client: Jenny

Short Desc: To find which method the best to predict CPU performance based on several hardware parameters of CPU. Goal is as consideration for a vendor in producing their hardware. For training, I use one of popular method which is K-Fold Cross Validation.

Calculate Number of Potatoes (Yellowish Object) in Digital Images

Topic: "Digital Image Processing"
Client: CarrierGuide15

Short Desc:
This is an easy and simple project to localize and calculate number of potatoes in image(s) automatically. All aspects need to be considered before designing appropriate method. This is unfiltered natural image taken in uncontrolled circumstances, we can image how unpredictable was the brightness and contrast present in whole part of images. So, I designed which best method to overcome this problem. Result and image data set are depicted by figures below.

Content Based Image Retrieval (CBIR) of Natural Images

Topic: "Digital Image Processing"
Client: Mrs. Sheila

Brief Exp:

This project was intended to retrieve image(s) from large image database based on its content. Features extraction was developed based on 2-D level-1 wavelet filtering toward image texture and colors. Training was developed based on ensemble training to multi layer neural network. This kind of training method is new for me and quite interesting to explore. I found that its quite good to be implemented to other cases. Result shows that feature extracted for this project could fetch similar images based on one image as an input. Could be used as basis to next development as well as adding improvement to it.

Thursday, March 6, 2014

Simple Shortest Path Finding Using Djikstra

Topic: Computation

This project was intended to make simulation of Djikstra path finding.

Histogram Equalization Transform Applied to Various Type of Contrast

Topik: Digital Image Processing
Mr. Enas

This project was intended to apply certain histogram equalization method to enhance various contrast type in digital images.


Wells Temperature Location Detection

Topic: Digital Image Processing
Jans Hendry

This application was intended to develop a program to detect wells location. These well indicate temperature value. Wells have circle shape so by detecting the circle, we can segment the image to get only wells with temperature.

Canny Edge Detector Improvement

Topic: Digital Image Processing

This project was intended to develop a canny improvement algorithm by involving positive and negative left diagonal and extend complexity time become 2N^2. Canny itself has N^2 complexity time.

Thursday, January 30, 2014

Contrast Image Enhancement on Low Light Using Opening by Reconstruction

Topic: Digital Image Processing
Client: Careerguide15

Motivation: To enhance low light image contrast using morphological methods based on paper "Morphological Background Detection and Enhancement of Images With Poor Lighting"

Thursday, December 12, 2013

Monday, September 23, 2013

WEBCAM Based DIY Laser Rangefinder

Topic: WEBCAM Based DIY Laser Rangefinder
Mr. Atul Yadav
Jans Hendry

Note: This program is intended to measure distance of object from camera using WEBCAM and Laser Pointer. Final task is to measure size of an object such as human height and other object.

Backtesting Data of Price List

Theme : Economy
Topic: Backtesting Data of Price List
Mr. Ray Prasad (e*ance)
Jans Hendry

Offline Signature Identification Using ANN-BP

Theme : Digital Image Processing
Topic: Offline Signature Identification Using ANN-BP
Barry Caesar Oktariyadi
Jans Hendry