Hi，Thanks for your sharing!
I want to change a color optical flow image(3 channels) to .flo files ,so that i can evaluate it with my own .flo
files,Did you write the function before?Thank you very much!

Hi, thanks for your contribution, but I have a question about how can I calculate the dense optical flows between two different Image frames, or this script only used for optical flow results' visualization? Thanks a lot!

Hello, ruoteng. Recently, I'm trying to reproduce Flownet2 in tensorflow. So, the coding implementation of AAE(average angular error) and AEE(average endpoint error) are what I need.I forked your rep(OpticalFlowTookkit) cause it is awesome and helpful, but there are some questions confused me at line 181 in flowlib.py.

def flow_error(tu, tv, u, v):
    """
    Calculate average end point error
    :param tu: ground-truth horizontal flow map
    :param tv: ground-truth vertical flow map
    :param u:  estimated horizontal flow map
    :param v:  estimated vertical flow map

Now, I have a numpy array(named ground_truth) converted from ground truth .flo file, which has shape of [1024, 436, 2].So, the param tu equals ground_truth[ : , : , 0] and param tv equals ground_truth[ : , : , 1], Is my understanding right? And can I get AAE by canceling the annotation in function flow_error?
what's more, I have an implementation by myself, I ALSO have no idea if it's understanding right.

def get_AEE(labels, predictions):
    num_sample, height, width , _= predictions.shape.as_list()
    squared_difference = tf.square(tf.subtract(predictions, labels))
    #此处axis很重要，有batch的四维数组的axis必须是 3 ， 三维数组得是 2
    loss = tf.reduce_sum(squared_difference, 3, keep_dims=True)
    loss = tf.sqrt(loss)
    return tf.reduce_sum(loss) / (num_sample*height*width)

Thanks very much! Any suggestion will help.

Hi,
I was planning to develop an optical flow library for Python.
I was wondering if you don't mind that I use your flowlib.py file to do so. If you agree, could you license the file?

Thanks in advance

data/prepare_train_data.py 的94行是不是写错了？dump_example函数的参数没有传进去吧？
Parallel(n_jobs=args.num_threads)(delayed(dump_example)(n) for n in range(data_loader.num_train))

When I use the toolkit to visualize the result of flow, the result is not good. Could you give me some advice?

如题

There's a type error between image and flow when in warping warp_image line 395

TypeError: Cannot cast ufunc add output from dtype('float64') to dtype('int64') with casting rule 'same_kind'

Can fix by adding fl = fl.astype('int64')

There is no module named png in the flowlib of lib.

Hi, @liruoteng
thank you for your sharing, I installed OpticalFlowToolkit according to your instructions. I successfully visualized the. flo file and converted it to png file, but I encountered a problem, I found the lower part of the picture was covered. I am not sure what this is. Can you help me?

I0627 14:53:11.230944 24143 net.cpp:460] conv1 <- input
I0627 14:53:11.230948 24143 net.cpp:417] conv1 -> conv1
*** Aborted at 1498546391 (unix time) try "date -d @1498546391" if you are using GNU date ***
PC: @ 0x7f67d21d704d caffe::CuDNNConvolutionLayer<>::LayerSetUp()
*** SIGFPE (@0x7f67d21d704d) received by PID 24143 (TID 0x7f67d2eda740) from PID 18446744072939728973; stack trace: ***
@ 0x7f67d05c54b0 (unknown)
@ 0x7f67d21d704d caffe::CuDNNConvolutionLayer<>::LayerSetUp()
@ 0x7f67d212999e caffe::Net<>::Init()
@ 0x7f67d212ba63 caffe::Net<>::Net()
@ 0x408be7 test()
@ 0x4072f0 main
@ 0x7f67d05b0830 __libc_start_main
@ 0x407a59 _start
@ 0x0 (unknown)

The resulting FLOW is stored in flownets-pred-NNNNNNN.flo

I have run make -j 5 all tools successfully, but when I run ./demo_flownet.py S data/0000000-img0.ppm data/0000000-img1.ppm, this error happened, and there is no .flo result. How can I solve it?

Hi liruoteng!
I'm currently trying to display the .flo files.
To do so i went throught your nice toolikt, got it setup on my desk and unfortunately I get the following error while launching the "python demo.py" :

Traceback (most recent call last):
File "demo.py", line 12, in
flow = kittitool.flow_read(flow_file)
AttributeError: 'module' object has no attribute 'flow_read'

Thanks in advance for your support :D !
Cheers

I would like to use this software as part of an existing project in python 3. Is this likely to work without significant modifications?

Thank you in advance.

The current repo does not seem to have the KITTI example, only the MIddlebury folder, under data/example.

demo.py tries to load in a PNG file instead of a .flow file for KITTI.