问题描述
我已经使用 Tensorflow 对象检测 API(ssd_mobilenet_v1 模型)训练了一个具有自定义数据集(加菲猫图像)的模型，并在 Tensorflow 存储库上提供的 android 示例应用程序中引用了它.该应用程序只能检测到大约小于或等于 20 厘米的距离的图像.
I've trained a model with a custom dataset (Garfield images) with Tensorflow Object Detection API (ssd_mobilenet_v1 model) and referring it in the android sample application available on Tensorflow repository. The application can only detected the images in distances less or equal 20cm approximately.
你知道我可以改进模型以在更远的距离(大约 30 厘米或更远)执行识别吗?
Do you have any clue about I can improve the model to perform recognitions in longer distances (about 30cm or more) ?
我不知道这个限制与我使用的输入大小有关(用 300x300 和 68x68 的图像测试)或需要任何自定义数据增强来改进它.
I don't know with this limitation is related with input size I'm using (tested with images with 300x300 and 68x68) or any custom data augmentation is needed to improve that.
推荐答案
众所周知，SSD 模型在小对象上的性能较差.您是否尝试过使用我们的 FasterRCNN 模型之一来查看结果是否可以接受?
SSD models are known to have worse performance on small objects. Have you tried using one of our FasterRCNN models to see if the result is acceptable?
这篇关于带有自定义数据的 SSD mobilenet 训练模型仅识别短距离图像的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！