top of page

Eve Smile May 2026

# Smile intensity (mouth opening + lip corner pull) mouth_width = distance(left_mouth, right_mouth) mouth_height = distance(upper_lip, lower_lip) intensity = min(100, (mouth_width / normalized_width) * 50 + (mouth_height / normalized_height) * 50)

Future<void> loadModel() async _interpreter = await Interpreter.fromAsset('smile_model.tflite'); eve smile

# Symmetry (difference between left and right smile pull) left_cheek = face_landmarks.landmark[234] # left cheek right_cheek = face_landmarks.landmark[454] # right cheek symmetry = 100 - abs(left_cheek.y - right_cheek.y) * 200 # Smile intensity (mouth opening + lip corner

# Duchenne marker (eye squint) left_eye_open = eye_aspect_ratio(face_landmarks, is_left=True) right_eye_open = eye_aspect_ratio(face_landmarks, is_left=False) duchenne = 1 if (left_eye_open < 0.25 and right_eye_open < 0.25) else 0 right_mouth) mouth_height = distance(upper_lip

Future<double> detectSmile(CameraImage image) async // Convert CameraImage to tensor input (224x224 RGB) var input = preprocessImage(image); var output = List.filled(1, 0).reshape([1, 1]); // output: smile score 0-1

_interpreter?.run(input, output); return output[0][0] * 100;

1. Product Overview EVE Smile is a mobile-first application that uses computer vision, voice analysis, and positive psychology to help users improve emotional well-being through guided smile exercises, mood tracking, and real-time feedback.

Subscribe to be notified about
new and upcoming events!

Thanks for subscribing!

Contact

Department of Informatics and Networked Systems

School of Computing and Information

University of Pittsburgh

135 N. Bellefield Avenue

622 IS Building

Pittsburgh, PA  15260

​​

Tel: (412) 383-4641

E-mail: ​[email protected]

  • Black LinkedIn Icon
  • Black Twitter Icon
  • Black YouTube Icon
  • GitHub-Mark-120px-plus

%!s(int=2026) © %!d(string=Elegant Pulse).

SEND ME A QUICK MESSAGE

Thanks for submitting!

bottom of page