Image Recognition

Image Recognition Domain

Image Recognition problems cover a large domain:

• some applications are currently deployed in production;
• some applications are currently under development;
• some applications are definitely feasible with some R&D work.
• some ideas are not feasible.

Image Recognition Classification & Attrasoft Domain

Classification:

Attrasoft classifies Image Recognition into 5 categories:

Category 1: A Newly captured image is a scanned image of an object in the master library.

• (examples: documents, magazines, passport, books...)

Category 2:  A Newly captured image is obtained by a camera. The condition is a fixed condition (distance, angle, light, ...). The object is in the master library.

• (examples: wheel rims, auto-parts, products, coins, antiques, pictures, paintings…)

Category 3: A Newly captured image is obtained by a camera. The condition has small and fixed variations (distance, angle, light, .. ). The object is in the master library.

• (examples: wheel rims, auto-parts, face, animals, salamanders…)

Category 4: A Newly captured image is obtained by a camera. The object is moving linearly. There are restrictions in the object orientation. The object is in the master library.

• (examples: aerial surveillance, counting, tracking, security camera, individual-based advertisement, wheel rims, auto-parts, face, animals, salamanders …)

Category 5:  All other cases.

Feasibility & Attrasoft Domain:

Category 1: Attrasoft has Category 1 applications currently deployed in production lines.

• A feasibility project will demonstrate requirements are achievable.

Category 2 is definitely feasible.

• A feasibility project will demonstrate requirements are achievable.

Category 3 is feasible.

• A feasibility project will determine conditions under which the project requirements are achievable.

Category 4 requires extensive R&D.

• A feasibility project will determine conditions under which the project requirements are achievable.

Category 5: Attrasoft generally does not accept projects in Category 5.

Pricing Model

Fixed Price Project Model 

• This model is used for the Feasibility / Proof-of-Concept Phase.

• With our Fixed Price Projects, ATTRASOFT offers you a cost effective & low risk method of developing software applications foryour organization.

• This model is used for projects where the client has a clearidea about the requirements of the project.

• ATTRASOFT helps the client define the best approach within the AME (Attrasoft Matching Engine) scope for developing theircustomized image recognition & data mining applications,

• including writing 80% of the SOW. 

• ATTRASOFT uses its experience / expertise with image recognition technology to deliver a high quality solution on time. 

Time & Material Project Model 

• ATTRASOFT can also execute a Time & Material Based Project Model for your organization.

• This model is suitable for projects that have very dynamic requirements.

Examples are: 

• The development of a new & revolutionary product where design & implementation are:

• constantly changing to adapt to new requirements; 

• only visible to the current phase due to the R&D nature;

• Mission critical projects where a detailed system specification cannot be created or provided by the client due to time constraints.

• ATTRASOFT can provide Services for these projects using the Time & Material Model with very aggressive implementations. 

Phased Fixed Price Project Model 

• Because Image Recognition & Data Mining are on the front line of technology, most clients prefer a phased approach.

• In this model some phases are done on a Time & Material Basis Model, while others are done on a Fixed Price Model.

• This model uses the Time & Material Model for certain phases, which are dynamic in nature ; & 

• the Fixed Price Model for certain phases, which have a defined specification. 

 

Image Recognition Procedure

(1) Set up the system (you can do that);

(2) Collect the data (you can do that);

(3) Preliminary Assessment via our off-the-shelf software, the ImageFinder.

• You should get Identification Rates ranging from 60% to 89%.
• The best rate, one of our customers (without any customization) was able to obtain, was an 89% Identification Rate.

• Note: The off-the-shelf ImageFinder has 70 open parameters for users to adjust

• which is the reason customers are only able to achieve Identification Rates ranging from 60% to 89%. 

• Note: The ImageFinder itself has 3,000+ internal parameters which users have no access to;

• customization is the process of adjusting these 3,000+ internal parameters for your specific image type;

• this is where the high degree of accuracy for your specific image type is obtained.

 (4) Feasibility Demo Project via Customized Stand-alone Software.

• ATTRASOFT will develop a stand-alone software, which will address the special needs / requirements of your application.

• This will allow you to show your upper management the practicality of a larger project.

(5) Programming Library in .Net;

(6) System Integration to your Production Line;

(7) Licensing & Annual Support.

• One-time License:         2% of the project.
• Annual License:             0.6% of the project
• Annual Support:             20% of the project. (First                                                year will be required.)