SMART ATTEND AI

Inspiration

Traditional attendance systems often fail to understand the real situations happening in a college. Students participating in hackathons, sports events, industrial visits, and other official activities may have approved On-Duty (OD), but camera-based systems can still mark them absent. Approved leave, late entry, unexpected timetable changes, and substitute faculty assignments also create difficulties in attendance management.

We were inspired to develop SMART ATTEND AI, an intelligent campus attendance ecosystem that goes beyond simply detecting faces. Our vision is to create a system that understands student attendance in context, adapts to timetable and faculty changes, and makes attendance management more accurate, flexible, and transparent.

What it does

SMART ATTEND AI is an AI-powered smart attendance and campus coordination system that connects students, faculty, administrators, timetables, and attendance records in one platform.

Blink-to-Start Smart Attendance

Faculty can initiate the attendance process using a predefined eye-blink gesture. Once activated, the classroom camera scans the room and uses face recognition and liveness detection to identify students and record attendance. The system is designed to avoid unnecessary continuous camera monitoring.

Intelligent OD and Leave Management

Students can apply for OD or leave through the application by providing the date, period, reason, and supporting documents. AI assists in analyzing and categorizing requests, such as sports activities, technical events, hackathons, and medical leave. The authorized teacher reviews and approves or rejects each request. Approved OD and leave records are automatically considered during attendance processing.

Context-Aware Attendance Reconciliation

The system does not simply treat every undetected face as an absence. It checks the attendance result against approved OD, leave, and relevant late-entry records before determining the final status.

The system supports Present, Absent, OD, Leave, and Late Entry statuses. Uncertain cases can be flagged for faculty review rather than being automatically classified incorrectly.

Dynamic Timetable Synchronization

Attendance is associated with the correct subject and period according to the latest college timetable. If a period, subject, or schedule changes, students receive notifications about the updated timetable. Attendance is then recorded according to the revised schedule.

Flexible Faculty and Substitute Management

Faculty assignments can change according to college requirements and staff availability. When a substitute teacher is assigned to a period, the system provides authorized access to record attendance for that specific class. Access can be limited to the assigned period and automatically expire afterward.

Role-Based Access Control

The system provides separate access for students, regular faculty, substitute faculty, and administrators. Students can view attendance and submit requests; faculty can manage attendance for authorized periods; and administrators can manage timetables, faculty assignments, permissions, and reports.

Notifications, Reports, and Attendance Disputes

The platform provides timetable-change notifications, OD and leave application updates, attendance history, attendance summaries, and low-attendance alerts. Students can also raise attendance disputes when their recorded status appears incorrect, allowing authorized faculty to review and correct records.

How we built it

We designed SMART ATTEND AI as a modular web-based platform with separate student, faculty, substitute faculty, and administrator dashboards.

The system combines a timetable management module, attendance processing module, OD and leave management module, faculty assignment module, notification system, and attendance reconciliation engine.

The planned workflow is:

The system identifies the current period from the latest timetable.
The assigned faculty member initiates attendance using the predefined blink gesture.
The camera captures the classroom view and processes student faces.
The attendance engine checks recognition results against approved OD, leave, and late-entry information.
The system generates the attendance status and flags uncertain cases for review.
Attendance records are stored and made available to authorized users.
Timetable changes and faculty substitutions update the relevant assignments and notifications.

AI and computer vision are intended to support face recognition, liveness detection, and request classification. Database management maintains student details, attendance records, applications, timetable information, and faculty permissions. Role-based access control protects attendance records from unauthorized changes.

Challenges we ran into

One major challenge was handling cases in which a student is not detected by the classroom camera but has a valid reason for being away, such as approved OD or leave. We needed to design a process that considers both camera results and administrative records.

Another challenge was ensuring that attendance always corresponds to the correct period, especially when the timetable changes unexpectedly. Managing substitute faculty access also requires careful permission control so that only authorized staff can modify attendance.

Other challenges include face recognition under different lighting conditions, detecting multiple students in a classroom, distinguishing intentional blink gestures from natural blinking, protecting sensitive student information, and maintaining accurate records across multiple modules.

Accomplishments that we're proud of

We are proud of developing a unified solution that addresses several connected college attendance problems instead of focusing only on face recognition.

Our main contributions include:

Blink-triggered classroom attendance.
AI-assisted face recognition and liveness detection.
OD and leave applications with teacher approval.
Context-aware attendance reconciliation.
Dynamic timetable synchronization and notifications.
Flexible faculty and substitute assignments.
Temporary, role-based attendance permissions.
Late-entry recording and attendance dispute handling.
Separate dashboards for students, faculty, and administrators.
Centralized attendance history and reporting.

Our key design principle is that an undetected face should not automatically mean an unjustified absence. The system considers approved information and flags uncertain cases for review, making attendance management more context-aware and transparent.

What we learned

Through this project, we learned that effective AI solutions must address real-world workflows rather than automate only one task. Attendance management involves interconnected processes, including classroom recognition, timetables, faculty availability, OD, leave, notifications, and permissions.

We explored how computer vision, AI-assisted classification, database management, access control, and real-time notifications can work together. We also recognized the importance of privacy, human approval, data accuracy, and manual review when automated results are uncertain.

What's next for SMART ATTEND AI

Our next step is to develop and test a functional prototype, starting with the student and faculty dashboards, timetable synchronization, OD and leave approval, and attendance reconciliation. We then plan to integrate and evaluate the real camera-based recognition and blink-triggered activation features.

Future improvements include integration with existing college management systems, better performance in large classrooms, offline attendance processing, multilingual support, advanced attendance analytics, predictive low-attendance alerts, and stronger privacy protections.

Our long-term vision is to transform SMART ATTEND AI into a complete intelligent campus platform where attendance adapts to real classroom conditions, changing schedules, faculty assignments, and students' approved academic activities.