PRESENTED BY:
Narmadha.K
Panimalar.R
Pavithra.P
Praveena.M

---

[attachment=11222]
FLEXIBLE HARDWARE ARCHITECTURE OF HIERARCHICAL K-MEANS CLUSTERING FOR LARGE CLUSTER NUMBER

ABSTRACT
 To Handle Large Cluster Number In Embedded Systems, A Hardware Architecture Of Hierarchical K-means (Hk-means) Is Proposed To Support A Maximum Cluster Number Of 1024.
 It Adopts 10 Processing Elements For The Euclidean Distance Computations And The Level-order Binary-tree Traversal
 The Gate Count Of The Hardware Is 414 K, And The Maximum Frequency Achieves 333 Mhz.
 It Supports The Highest Cluster Number And Has The Most Flexible Specifications Among Our Works And Related Works
Tools required
DEVICE
◦ FPGA Spartan3
SOFTWARE
◦ Xilinx ISE 8.1i
◦ Xilinx Platform Studio
◦ Visual Basic
◦ Matlab
LANGUAGE USED
◦ VHDL & C
Tools required
DEVICE
◦ FPGA Spartan3
SOFTWARE
◦ Xilinx ISE 8.1i
◦ Xilinx Platform Studio
◦ Visual Basic
◦ Matlab
LANGUAGE USED
◦ VHDL & C
Objective
◦ Most sensing applications require some form of digital signal processing. This processing can be performed on an FPGA rather than a microprocessor or DSP.
◦ To handle large cluster number in embedded systems, a hardware architecture of hierarchical K-Means (HK-Means) is proposed.
Existing System
◦ The costs in embedded systems are always important, and the struggle between the computational time and the hardware area becomes severe especially as the cluster number increases. the large cluster number is still a design challenge for K-Means hardware architectures.
Proposed System
◦ New architecture includes a hierarchical memory structure to store the cluster centroids for distance calculations and binary-tree traversal are employed to compute the nearest centroid operations in pipeline.
K-means clustering
◦ Clustering is a way to separate groups of objects.
◦ It finds partitions such that objects within each cluster are as close to each other as possible, and as far from objects in other clusters as possible.
◦ K-means clustering requires that you specify the number of clusters to be partitioned and a distance metric to quantify how close two objects are to each other.
Flow chart
◦ Matlab GUI (Graphical user interface)
◦ MATLAB implements GUIs as figure windows containing various styles of unicontrol objects.
◦ GUI includes laying out the components, programming them to do specific things in response to user actions.
◦ GUIDE also generates an M-file that contains code to handle the initialization and launching of the GUI.This M-file provides a framework for the implementation of the callbacks the functions that execute when users activate components in the GUI.
GUI file
M-FILE
GUI OUTPUT – TEXT FILE CREATED
Advantages
◦ With a large number of variables, HK-means may be computationally faster.
◦ Hierarchical clustering has the distinct advantage that any valid measure of distance can be used.
Applications
◦ To locate tumors
◦ Measure tissue volumes
◦ Diagnosis
◦ Fingerprint recognition
◦ Video segmentation
◦ Color quantization

Hey, I have to do the same project using (spartan 3E or Spartan6 )

I need to send image features (1158 samples x 21 features) to Spartan 3e FPGA board from the matlab workspace with rs232 and then do K-means classification on that and then return the data (1158samples x 1 group(1-4)) from the FPGA back into the matlab workspace. How do I do that? Kindly let me know if there is some tutorial which is spartan 3e specific on this which can tell me how to do it step by step.

Thanks and Regards,
Cavalry

I want this complete project with code. if u can provide it for me i will pay money for u. reply for my mail. star.raj453[at]gmail.com

hi...
have u implemented any application for that HK means ...
coz am trying to implement that "locating tumour"...
if it coincides pls help on this