SRS Document
The purpose of this document is to present the simulation known as iProVis: Intelligent Product Vision System. This artificial intelligence-based vision system is designed to provide users with the ability to easily identify the cheapest products and locate the nearest availability of these products. The document provides detailed information about the project requirements, addressing identified constraints, and the proposed software functionalities. It also outlines the software technologies utilized in the project’s execution, highlighting the positive impacts these technologies have had on overall system performance. Additionally, this Software Requirements Specification (SRS) defines how participants interact with the model-based computer vision system.
Today, customers have numerous shopping options, including online and physical stores, each with various shopping methods. The rapid growth of online shopping has made life easier for customers, thanks to software solutions tailored for this space. However, this convenience has also led to challenges such as "getting away from the product itself," with numerous sellers offering the same product. As a result, while e-commerce platforms offer various opportunities for sellers, the process of choosing the best product at the most affordable price has become increasingly difficult for the end user. The price differences across sellers on multiple platforms are often minimal, which complicates the decision-making process.
Despite the advancements made by developed online marketplaces, customers still struggle with finding the "most preferred product at the most affordable price." This challenge becomes more pronounced when shopping in physical stores. Our system is designed to address these difficulties. With our solution, customers can easily determine the best price and the nearest store location for a product they are considering. If the product is unavailable in a physical store, the system will recommend the most suitable online marketplace. The key insight behind this system is that customers instinctively seek “better prices” when they find products in physical stores that seem overpriced. Our AI-powered software, equipped with advanced artificial intelligence models, will provide real-time information on the best prices and locations for these products, ensuring that end users can make purchases under the most optimal conditions.
- End User: The person or group who uses a product or service at the final stage.
- E-commerce: The buying and selling of products or services over the internet.
- Physical Stores: Brick-and-mortar locations where products are sold directly to customers
- face-to-face.
- Artificial Intelligence (AI): Computer systems and software designed to mimic human
- intelligence.
- AI Model: An algorithm trained to perform a specific task or make predictions.
- Smartphone: A mobile device that combines phone functions with advanced features like
- internet access and application use.
- API (Application Programming Interface) : Protocols for software interaction, enabling
- communication between systems.
- Camera Interface : Feature allowing users to capture product images for information
- retrieval.
- Database : Structured data storage for user profiles, product data, and transactions.
- Hardware Interfaces : Physical components the application interacts with, like cameras and
- sensors.
The second section of this document describes the functionalities of iProVis: Intelligent Product Vision System. The Overall Description provides an overview of the project's objectives, areas of application, and its intended purpose. It explains in detail how the project will assist customers and highlights the features that underscore its importance for the end user. One of the key sections, Development Methodology, further explains the project's construction process and the methods used to guide its development.
This project aims to revolutionize the retail experience by leveraging computer vision and deep learning technologies to replace traditional barcodes and QR codes for product recognition in sectors such as supermarkets, clothing, food, and electronics. The developed system addresses multiple consumer needs, including effortless access to product price information, comparative price analysis across stores, and multilingual support for accessing detailed product content information.
The product is designed as a standalone mobile application integrated with the iPRoVis system's API for advanced functionality. This integration enables not only product identification but also facilitates an intelligent recycling process by recognizing recyclable packaging types, determining whether the packaging is empty or full, and providing weight details. This functionality ensures efficient sorting and recognition by smart recycling machines, thereby reducing the time and effort required for waste disposal and recycling.
The system is developed with the dual purpose of enhancing the consumer shopping experience and contributing to sustainability goals under a green IT strategy. By promoting environmentally friendly practices and providing advanced features like multilingual support, the product caters to a diverse user base from various cultural and linguistic backgrounds. The project also proposes an innovative framework for recycling by connecting product recognition capabilities to smart recycling systems, encouraging sustainable behaviors among users. This holistic approach not only solves immediate user needs but also aligns with long-term sustainability objectives, making it a transformative solution for both the retail and environmental sectors.
For the development of this project, we have adopted the Agile methodology, specifically leveraging the Scrum framework, which is ideal for dynamic and iterative project needs. Scrum divides the work into smaller, manageable cycles called sprints, each lasting a fixed duration of 2-4 weeks and delivering a potentially shippable product increment. This iterative approach enables flexibility and continuous improvement throughout the development process. In our Scrum-based workflow, tasks are organized and tracked through a Scrum Board, which consists of six distinct phases: Project Backlog, containing all planned tasks; To Do, listing prioritized tasks ready for the sprint; In Progress, showing tasks currently under development; In Review, for tasks being tested and validated; To Deploy, indicating modules ready for integration; and Done, where completed and verified tasks are marked. These phases ensure a structured and transparent workflow.
Each sprint begins with detailed planning, including story points and risk assessments for tasks. The team conducts daily stand-up meetings, limited to 15 minutes, to communicate progress, address potential blockers, and ensure alignment toward sprint goals. Scrum roles are clearly defined to streamline this process: the Product Owner defines and prioritizes project requirements, the Scrum Master ensures adherence to the Scrum framework and facilitates team efficiency, and the Development Team collaboratively works to achieve sprint objectives. Scrum's iterative nature allows the team to adapt to evolving requirements and incorporate stakeholder feedback after each sprint. Delivering a working increment at the end of every sprint ensures early validation and continuous improvement of the product. This adaptability is particularly advantageous for the project, as it accommodates changes while maintaining high- quality outcomes.
This project aims to transform the retail industry by replacing traditional barcode and QR code scanning with computer vision technology for product recognition. Through its mobile application, users can access product price information, compare prices across stores, and obtain multilingual product details. Additionally, the system integrates with smart recycling through the iPRoVis API, supporting sustainability by enabling automated sorting of recyclable packaging. By incorporating Agile principles and leveraging Scrum, this project ensures a structured yet flexible approach. The methodology facilitates timely, high-quality delivery while addressing innovative and sustainable goals, ultimately enhancing consumer experiences and contributing to environmental responsibility.
This section defines the distinct user groups who will interact with the system. Users are classified based on their technical knowledge, expectations, and roles, providing clarity on how the system is designed to meet their needs.
Participants represent the general consumer audience who use the mobile application to access product information. Their interaction with the system includes learning about product prices, comparing prices across stores, and accessing detailed product content information.
Participants are typical consumers who rely on the system to make informed purchasing decisions by comparing prices and analyzing product details through the mobile application.
Participants require no prior technical knowledge to use the system. A user- friendly interface ensures they can navigate and complete tasks effortlessly.
Participants can benefit from multilingual features that enable them to view product information in multiple languages, ensuring accessibility across diverse regions.
The system supports participants in streamlining recycling processes by integrating with smart recycling systems for products designed with sustainable practices.
Admins are the authorized personnel responsible for managing and overseeing the system’s functionality. Their primary roles include updating product data, maintaining system integrations, and ensuring overall performance.
Admins monitor and manage the system’s operations, ensuring smooth functionality and addressing user feedback effectively.
Admins are responsible for updating product information, including pricing and descriptions, and addressing user-submitted feedback.
Admins require basic computer skills to operate the system and perform their duties efficiently.
Admins oversee the integration of the iPRoVis API and smart recycling systems, ensuring seamless functionality and accurate data flow.
Admins must ensure the accuracy and consistency of content across multiple languages and may require familiarity with different languages to manage multilingual features effectively.
Mobile Application Interface
- The design should be user-friendly, enabling smooth navigation.
- A camera interface should be integrated for product image capture.
- Display detailed product information, including price, stock status, and multilingual
- content.
- Include a price comparison feature that shows a list of stores and their prices.
- Provide a user profile management section to save preferences and viewing history.
Camera
- Integration with the device's camera for image capture.
Sensors
- Utilize ambient light sensors to optimize image capture conditions.
Internet Connectivity
- Support for Wi-Fi or mobile data connections to facilitate server communication.
Operating Systems
- The app should be compatible with Android platforms.
APIs
- Integration with third-party APIs for price comparison and retrieving product information.
Database
- The application should connect to a backend database to store user profiles, product data, and transaction history.
Network Protocols
- HTTPS should be used for secure communication with the iProViS server.
Data Formats *JSON or CSV formats should be employed for data exchange between the mobile app and the server.
Primary Actor: Customer
Goal: Allow users to log into their account or create a new one for personalized features.
Preconditions: The app is installed, and the user is either logging in for the first time or after a session has ended.
Postconditions: The user gains access to personalized features, including saved searches and activity history.
Main Flow:
1) Open the app: The user launches the app for the first time or after logging out.
2) Login/Register prompt: The app offers options to log in or create a new account (e.g., “Log In” or “Create Account”).
3) Login: If the user has an existing account, they enter their credentials (e.g., email or username and password). The app verifies these and grants access to personalized features.
4) Registration (for new users): When selecting “Create Account,” the user provides necessary details (email, password, name, shipping address, etc.). Once submitted, the account is created, and the user is logged in.
5) Account settings: Users can modify preferences like language or notification settings.
6) Access personalized features: After logging in, users can scan products, view saved preferences, and receive targeted offers.
Alternative Flow:
- Forgot Password: If the user forgets their password, a reset link is sent via email.
- Social Media Login: Users can log in via social media accounts (e.g., Google, Facebook, or Apple ID).
- Skip Login: Users can skip the login and use the app in limited mode, with an option to log in for a personalized experience.
Primary Actor: Customer
Goal: Capture a product photo to retrieve detailed information.
Preconditions: The app is installed, and the product is visible to the camera.
Postconditions: The user receives detailed information, including price, stock status, price comparisons, recyclability, and multilingual options.
Main Flow:
- The user opens the app and selects "Scan Product" or "Take Photo."
- The app prompts the user to take a clear photo of the product.
- The user captures the photo.
- The app uses computer vision to identify the product.
- Product details are displayed:
-Price: The product's current price.
-Stock Availability: Whether the product is in stock or not.
-Price Comparison: Comparisons across stores (online or offline).
-Recyclability: Information about whether the product or packaging is recyclable.
-Multilingual Support: Option to view details in different languages.
Alternative Flow:
-If the product cannot be recognized, the app notifies the user and offers manual entry or retry options.
Primary Actor: Customer
Goal: View price and stock status for a product.
Preconditions: The user has taken a product photo, and the app has recognized it.
Postconditions: The user views current pricing and stock information.
Main Flow:
-
After capturing the product photo, the app retrieves and displays the price.
-
It shows stock availability at nearby stores or online retailers.
-
The user can click on a store or retailer to view more details.
Alternative Flow:
-If the price or stock is unavailable, the app notifies the user and suggests alternatives or offers to notify them when the product is back in stock.
Primary Actor: Customer
Goal: Compare prices for the product across different retailers.
Preconditions: The user has taken a product photo, and the app has identified it.
Postconditions: The user sees a comparison of prices from various stores.
Main Flow:
-
After the product is recognized, the app retrieves price data from multiple retailers.
-
A comparison is displayed, showing the lowest and highest prices.
-
The user can select a retailer to view additional details.
Alternative Flow:
-If no comparison is available, the app notifies the user and offers an option for manual search.
Primary Actor: Customer
Goal: Display recyclability information about the product.
Preconditions: The user has taken a product photo, and the app has recognized it.
Postconditions: The user learns if the product or packaging is recyclable.
Main Flow:
-
After product recognition, the app displays recyclability information.
-
An icon or label indicates whether the product is recyclable (e.g., "Recyclable," "Not Recyclable," or "Check local guidelines").
-
If recyclable, the app provides additional recycling instructions.
Alternative Flow:
-If recyclability data is unavailable, the app prompts the user to check the packaging or contact the manufacturer for further information.
Primary Actor: Customer
Goal: View product information in the user’s preferred language.
Preconditions: The user has taken a product photo, and the app has identified it.
Postconditions: The product details are displayed in the selected language.
Main Flow:
-
Once product details (e.g., price, stock, recyclability) are shown, the app offers an option to switch to another language.
-
The user selects a preferred language (e.g., English, Spanish, French).
-
The app translates all details into the chosen language.
Alternative Flow:
-If the selected language is unavailable, the app informs the user or provides a translation tool.
Primary Actor: Customer
Goal: Access additional information, such as reviews and ratings.
Preconditions: The user has taken a product photo, and the app has recognized it.
Postconditions: The user gains insights into the product via reviews and ratings.
Main Flow:
-
After displaying product details, the app shows options to view customer reviews and ratings.
-
The app may display star ratings, verified reviews, and product demonstration videos.
-
The user can scroll through reviews or click on ratings for more details.
Alternative Flow:
-If reviews or ratings are unavailable, the app notifies the user and suggests similar products.
Primary Actor: Customer
Goal: Receive notifications about available deals or discounts.
Preconditions: The user has taken a product photo, and the app has recognized it.
Postconditions: The user is notified of any available deals.
Main Flow:
-
The app checks for any ongoing promotions on the product.
-
It displays any available deals (e.g., discounts, coupon codes, or bundle offers).
-
The user can click on the offer for more details or apply it to their purchase.
Alternative Flow:
-If no discounts are available, the app informs the user that there are no current offers.
Primary Actor: Customer
Goal: Assist users when the app fails to recognize a product.
Preconditions: The user has taken a product photo, but the app can't identify it.
Postconditions: The user is given an alternative method to find the product information.
Main Flow:
-
If the app fails to recognize the product, it notifies the user.
-
The user can manually enter the product name or retake the photo.
-
Other search methods (e.g., barcode scanning or uploading an image) are provided.
Alternative Flow:
-If manual search options fail, the app offers a "Help" feature.
The application must deliver an efficient, reliable, and responsive user experience, even under heavy usage. Here are the key performance considerations:
Product Recognition: The app should recognize a product within 2 seconds, assuming optimal lighting and focus conditions.
Price and Stock Retrieval: Product price and stock data should appear within 3 seconds of recognition.
Price Comparison: The comparison data should load within 5 seconds of retrieving product details.
Concurrent Users: The app must support at least 100 concurrent users without significant performance issues.
Data Handling: The backend should handle up to 10,000 transactions per hour without noticeable degradation.
Service Availability: The app must be available 99.9% of the time, with downtime only for planned maintenance.
Error Handling: The system should display appropriate error messages and automatically retry failed processes when needed.
User Growth: The backend should scale horizontally to accommodate growing user numbers and app traffic.
Database Scaling: The database must scale efficiently as the number of products, transactions, and user profiles increases.
Network Latency: Average network latency should remain under 1 second for most operations.
Image Processing Latency: Product recognition should occur within 2 seconds after image capture.
Memory Usage: The app should not exceed 200MB of RAM during normal operations.
Battery Consumption: The app should consume no more than 10% battery per hour during regular use.
Data Consumption: The app should use no more than 50MB of data per session for product searches and comparisons.
Server Load Balancing: Load balancing mechanisms must ensure even traffic distribution.
Failover Mechanism: Critical services should have a failover mechanism to maintain high availability.
3.3.8 Caching
Frequent Data Caching: Frequently requested data should be cached to improve performance. The cache should refresh at regular intervals to ensure up-to-date information.
These performance specifications are designed to ensure the app delivers a fast, reliable, and efficient experience under varying conditions, while being scalable, resilient, and responsive to user demands.
The app should be usable across various devices with different screen sizes and resolutions.
Product recognition should happen within 2 seconds under optimal conditions.
The app should support 100 concurrent users without degradation.
The app should have an intuitive interface with minimal learning required. Accessibility features must also be provided.
The app should adapt to different lighting conditions to ensure optimal image recognition.
The backend should be able to scale as the user base and product data grow.
The app must comply with data protection regulations (e.g., GDPR) and ensure user data privacy. Mechanisms should be in place to prevent unauthorized access to sensitive information.
[1] Li, Q., Peng, X., Cao, L., Du, W., Xing, H., Qiao, Y., & Peng, Q. (2020). Product image recognition with guidance learning and noisy supervision. Computer Vision and Image Understanding, 196, 102963.
[2] Bai, Y., Chen, Y., Yu, W., Wang, L., & Zhang, W. (2020). Products-10k: A large-scale product recognition dataset. arXiv preprint arXiv:2008.10545.
[3] Chen, S., Liu, D., Pu, Y., & Zhong, Y. (2022). Advances in deep learning-based image recognition of product packaging. Image and Vision Computing, 128, 104571.
PDF format of our SRS Document : iProViS_SRS.pdf