1
$\begingroup$

We are using 2D Laser Scanner to scan various objects of different geometric shapes for e.g. cylinder, spiked, cylinder with notch, cylinder with curved edges e.t.c. The dataset contains points in the format [x, y] with the dimension of 1 complete scan being 160x2. The goal is to use these scan points to classify the various shapes.

I have used a multilayer NN with sigmoid as the final layer and Adadelta optimizer for this problem but the accuracy reaches only upto 70%.

Can anyone recommend a proper model that can be used for Laser Scanner Data Classification?


MODEL

def baseline_model():
    model = Sequential()
    model.add(Dense(2048, input_dim=160, activation='relu'))
    model.add(Dropout(0.1))
    model.add(Dense(1024, activation='relu'))
    model.add(Dropout(0.1))
    model.add(Dense(512, activation='relu'))
    model.add(Dropout(0.2))
    model.add(Dense(256, activation='relu'))
    model.add(Dropout(0.3))
    model.add(Dense(128, activation='relu'))
    model.add(Dropout(0.4))
    model.add(Dense(64, activation='relu'))
    model.add(Dropout(0.5))
    model.add(Dense(32, activation='relu'))
    model.add(Dropout(0.5))
    model.add(Dense(6, activation='softmax'))
    Adam = optimizers.Adam(lr=0.001)
    Adadelta =  optimizers.Adadelta(lr = 1)
    model.compile(loss='categorical_crossentropy', optimizer=Adadelta,   metrics=['accuracy'])
$\endgroup$
  • $\begingroup$ You could convert this problem to shape detection on an image, where you can probably use few-layer convolution networks to solve the problem. You would introduce a lot of new variables but it would be easier to solve it that way. $\endgroup$ – Harsh Sinha Dec 8 '19 at 20:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.