Aplicación móvil híbrida utilizando visión computacional y redes neuronales para apoyar la resolución de casos de mascotas perdidas en la ciudad de Chiclayo
Cargando...
Fecha
2025
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Católica Santo Toribio de Mogrovejo
Resumen
Las mascotas están presentes en más del 60% de hogares en el país ya que juegan un papel crucial en el bienestar mental y físico de sus dueños. Sin embargo, la pérdida de estas mascotas causa fuertes emociones adversas en quienes las albergan, estimándose que alrededor de 420,000 mascotas se pierden anualmente en Perú debido a la eficacia limitada de los métodos de búsqueda convencionales y digitales. La "Asociación Refugio Animal Chiclayo", una entidad sin fines de lucro juega un papel crucial en la protección de animales en Chiclayo, abarcando actividades desde adopciones hasta reportes de mascotas perdidas. A pesar de sus esfuerzos, enfrenta desafíos significativos debido a la fragmentación en las redes
sociales y elevado volumen de mensajes, lo que dificulta la publicación y el seguimiento efectivo de los casos. Para abordar estas limitaciones, se propone desarrollar una aplicación móvil híbrida que utilice visión artificial y redes neuronales. Los objetivos específicos incluyen determinar el tipo de red neuronal apropiada, construir un modelo de red neuronal, obtener un nivel aceptable de precisión, construir la aplicación móvil para el despliegue y validar la usabilidad por parte de los miembros de la comunidad. La metodología empleada es Mobile D, que permite una iteración rápida y adaptación continua del desarrollo de la aplicación según las necesidades del proyecto. Tras el desarrollo de la aplicación, se logró un
modelo de detección y similitud altamente eficiente. Destaca particularmente la notable flexibilidad del segundo al brindar resultados precisos incluso con imágenes no utilizadas durante su entrenamiento.
Pets are present in over 60% of households in the country as they play a crucial role in the mental and physical well-being of their owners. However, the loss of these pets causes strong adverse emotions in those who care for them, with an estimated 420,000 pets being lost annually in Peru due to the limited effectiveness of conventional and digital search methods. The "Asociación Refugio Animal Chiclayo," a non-profit organization, plays a crucial role in animal protection in Chiclayo, encompassing activities ranging from adoptions to reports of lost pets. Despite their efforts, they face significant challenges due to fragmentation on social media and the high volume of messages, making it difficult to effectively publish and track cases. To address these limitations, the proposal is to develop a hybrid mobile application that utilizes computer vision and neural networks. Specific objectives include determining the appropriate type of neural network, building a neural network model, achieving an acceptable level of accuracy, constructing the mobile application for deployment, and validating usability by community members. The methodology employed is Mobile D, which allows for rapid iteration and continuous adaptation of application development according to project needs. Following the development of the application, a highly efficient detection and similarity model was achieved. The second model stands out particularly for its remarkable flexibility, providing accurate results even with images not used during its training.
Pets are present in over 60% of households in the country as they play a crucial role in the mental and physical well-being of their owners. However, the loss of these pets causes strong adverse emotions in those who care for them, with an estimated 420,000 pets being lost annually in Peru due to the limited effectiveness of conventional and digital search methods. The "Asociación Refugio Animal Chiclayo," a non-profit organization, plays a crucial role in animal protection in Chiclayo, encompassing activities ranging from adoptions to reports of lost pets. Despite their efforts, they face significant challenges due to fragmentation on social media and the high volume of messages, making it difficult to effectively publish and track cases. To address these limitations, the proposal is to develop a hybrid mobile application that utilizes computer vision and neural networks. Specific objectives include determining the appropriate type of neural network, building a neural network model, achieving an acceptable level of accuracy, constructing the mobile application for deployment, and validating usability by community members. The methodology employed is Mobile D, which allows for rapid iteration and continuous adaptation of application development according to project needs. Following the development of the application, a highly efficient detection and similarity model was achieved. The second model stands out particularly for its remarkable flexibility, providing accurate results even with images not used during its training.
Descripción
Palabras clave
Mascotas, Aplicaciones móviles, Visión artificial, Pets, Mobile applications, Computer vision
Citación
A. M. Serquen Manay. "Aplicación móvil híbrida utilizando visión computacional y redes neuronales para apoyar la resolución de casos de mascotas perdidas en la ciudad de Chiclayo," tesis de licenciatura, Fac. de Ingeniería, Univ. USAT, Chiclayo, Perú, 2025. [En línea]. Disponible en:
