The digital revolution was very good to the camera makers who jumped in early. In the film days prices for the very best cameras were well under $2000. The digital age boosted that to $8000, and decreased the product lifecycle of any given model to just a few years. But the revolution was ultimately good for photographers too. It was an exciting and very expensive time.
.
Camera maker's marketing departments are doing their best to keep the "revolution" going, but in fact digital cameras are now quite mature and the incremental developments are less revolutionary than they were in the past. For instance, the astoundingly low noise levels produced at 6-digit ISO settings are mostly due to in-camera massaging of files done after the raw sensor data is captured. In other words, to take advantage of this you have to put up with the severe limitations imposed by JPEG files. If you shoot RAW files to get away from those limitations you'll find the noise levels from new image sensors are modestly improved over those from a generation or two ago¹. Your current camera and some noise reduction software can probably do as well at settings below ISO 3200, and give you a far more useful TIFF file in return for the extra work. My point here is not to knock the new cameras, but rather to point out that the core technologies haven't changed as much as it seems lately.
.
Camera manufacturers are still making significant improvements, but many have less to do with the ultimate image quality a camera can produce and more to do with features that allow photographers to take advantage of the camera more easily. For example, "live view" can help to accurately and easily preview depth of field at the shooting aperture. It does not change what the camera is capable of in terms of image quality, but it helps prevent mistakes and the resulting ruined shots. Likewise, if your vision isn't what it used to be, a larger and better LCD screen can help you see histograms and settings more easily and quickly.
.
Some improvements may not do much for you. If the largest prints you intend to make are 12x18 inches and your camera has six or more megapixels, additional pixels won't help much. Conversely, it takes four times as many pixels to get the same resolution in a 24x36 inch image, so if large detailed prints are your thing, more pixels could help². A camera's frame rate could be significant if you shoot flying birds or fast action, or insignificant if you shoot only landscapes. Many new cameras create RAW files with 14 bits of information per color as opposed to the 12 bits of previous models, while JPEG files have only 8 bits per color regardless of the camera that creates them. More bits mean more malleable image files. The 12 bits per color produced by previous cameras are more than adequate most of the time, while 14 bit files can help minimize the banding and "blocking up" problems that sometimes occur when adjusting difficult images³. Video is related to still photography by shared digital technologies, not by art. The video capability in an SLR won't help your still images at all⁴.
.
Obviously, determining whether any camera improvement is significant depends on what you do with your images and how you use your camera. It's often the case that a new camera model will not produce image files that are visibly superior to your current camera, but it may let you produce the files you were after more easily. Taking stock of what you do, how you do it, and what might help, is simply a logical part of deciding whether that new camera is truly worthwhile.
.
.Happy shooting!

.
Notes:
.
¹ Noise levels on a pixel size basis have been reduced more than it would appear based on apparent image noise levels. Let me explain. The laws of physics tell us that, with all other things being equal, as pixels get smaller they become noisier. But all other things are not equal, and sensor technology improvements have made it possible to pack more smaller pixels on an image sensor and keep the noise levels the same, or even decrease them slightly. That means big improvements have actually been made, but they have been nullified by the increasing pixel counts. This tradeoff is mostly a marketing game. More pixels sell more cameras, even though other factors can be far more important in terms of image quality. With a little interpolation in Photoshop just 6 MP can make a remarkable 13x19 inch print if the image is not noisy. More pixels allow making larger prints, but few people do. For many, more flexibility in cropping is the only thing really gained by a higher pixel count. Also, a small format like 35mm can benefit from more pixels only to a certain point. Making a large print from the small 24mm x 36mm sensor results in extreme magnification of the small frame. Every tiny lens flaw and aberration is magnified as much That means it's important to use the best possible lenses on high resolution digital SLR cameras. But since diffraction sets theoretical limits as to how sharp any lens can possibly be at a given aperture, even a theoretically perfect lens has limits as to how much resolution it can make use of. In a nutshell, lenses of the same quality will always give better results on a larger format camera, and nothing can change this.
.
² This assumes printing on a smooth paper. Many use canvas as a way to make huge prints from files that would never support that size on paper. A given image can be printed much larger on canvas only because the fine details are obliterated by the texture. The image must become huge before the smallest details are large enough to be visible within the weave.  Many think photographic prints on canvas look like paintings. Others think they look like a prints viewed through a screen door. In the end it's a matter of personal taste.
.
³ Humans can discriminate shades to a maximum level of slightly less than 8 bits per color. That means banding in subtle gradations of color won't be seen in a completely un-manipulated (even by in-camera processing) 8-bit file. But if we apply a curve to increase contrast, banding will quickly become visible. Twelve bits per color is usually plenty, but increasing it to 14 bits gives more leeway for steeper curves and other manipulations without visible banding or "blocking up". Note that after a steep curve or other manipulation has been applied to a file with 12 or 14 bits per color, the file can then be converted to 8 bits per color without visible effect. But it's a one way street. The information that was lost in the conversion can never be regained.
.
⁴ Most of the components required for video were already present in digital SLR cameras, so it was cheap and easy to add video capability. Marketing departments sell video as a "must have feature" to bolster prices while adding almost nothing to the cost of producing a camera. That's reason enough for any company to do it. There's a great "cool factor" in video that helps sell these cameras, but few who have any real passion for creating still images will spend their time shooting, editing, and producing video instead. Still fewer will do it well, and even fewer will figure out what to do with the videos after they're made. It's worth pointing out here that large prints made from high resolution digital captures out-resolve video by a huge margin. That's another way of saying that we are a very long way from being able to display anything close to print resolutions on any LCD or other electronic display media. The best HD video has a total of only 1080 lines of vertical resolution spread over a large screen while a good print just 3 inches high will contain as much data.
.